Current estimates about the Covid-19 fatality rate may be too high
wsj.com
wsj.com
1) Extrapolating from the infection rates of very specific groups (for example, evacuees) to the entire population without taking into account transmission dynamics and the time between infection and detection does not make very much sense. The authors naively multiply the infection rate among evacuees by the population of Wuhan to conclude that Wuhan must have had 178,000 infections at the end of January. By comparison, epidemiological models have estimated there were around ~20,000 infections at that time [1][2][3]. What conclusion should we draw here? If you use sloppy, back of the napkin math to over-inflate the infection count by 10x then you can correspondingly deflate the mortality rate?
2) Speculating about the mortality rate of Covid-19 based on several gigantic assumptions ("If our surmise of six million cases is accurate, that’s a mortality rate of 0.01%") seems borderline irresponsible. Numerous researchers have been modeling this virus and have generally arrived at numbers in the range of 0.5% to 1.6% [4][5][6][7][8][9]. The authors don't present any compelling reason why we should doubt those numbers.
3) Ultimately the mortality rate is not as important a number as the hospitalization rate. The authors would have you believe this virus is no worse than the flu, but this is not congruent with the number of reports coming out of places like Italy and New York saying they're about to run out of ICU beds, or China rushing to build temporary hospitals to house all of the patients that need critical care. What the mortality rate might be under ideal circumstances where every patient receives adequate medical care might be significantly different compared to a scenario where you've run out of ICU beds and have to start rationing ventilators.
[1] https://www.mdpi.com/2077-0383/9/2/419/htm
[2] https://www.medrxiv.org/content/10.1101/2020.01.23.20018549v...
[3] https://www.mdpi.com/2077-0383/9/2/523/htm
[4] https://www.imperial.ac.uk/media/imperial-college/medicine/s...
[5] https://institutefordiseasemodeling.github.io/nCoV-public/an...
[6] https://cmmid.github.io/topics/covid19/severity/diamond_crui...
[7] https://www.medrxiv.org/content/10.1101/2020.03.04.20031104v...
[8] https://www.medrxiv.org/content/10.1101/2020.03.09.20033357v...
A person with flu and fever would probably stay home and just rough it out, a person with fever fearing the coronavirus will probably rush to the hospital.
If everyone having a fever in a flu season would go to the hospital and would demand medical attention we would be running out of hospital beds in normal scenarios as well.
They don't want you at the hospital unless it is absolutely necessary. Got a fever and still able te breath: stay at home.
It doesn't really matter whether the odds are more like winning a bottle of wine in a charity raffle or losing a game of Russian roulette.
You don't get admitted to hospital because you ask. You get admitted to hospital because a doctor admits you. Currently even covid-19 patients don't get admitted to hospital by asking. They need to be pretty severely ill to get admitted to hospital. Health care professionals are triaging as much as possible, and sending people who are suspected (or even confirmed) to have covid-19 back home unless the patient really needs to be admitted.
Despite all this triage hospitals are over run and many countries are experimenting with field hospitals built in large sporting gym halls or convention centres.
One thing that doctors are telling us is that covid-19 is not like flu for their patients.
You don't get a hospital bed just because you ask for it, you get one if a Doctor thinks you need one.
So your premise is inaccurate.
Next day I read a howling butthurt editorial in the WSJ with my morning coffee. Basically accusing the FDA of murder.
Since then I don't take the WSJ at face value. The only thing they care about is making money on corporate stocks.
Same thing is going on here. The WSJ only cares that they've taken a bath.
I 100% agree that we should be running antibody tests to see how many people have caught COVID-19 and recovered. All of our testing until recently was based on RNA tests, which may only be positive during a relatively short window. (I saw a case study claiming that even some hospitalized patients are testing negative inside 7 days.) We need to know how many people catch this and beat it quickly.
That said, even in populations that have been RNA-tested early and extensively (such as the Diamond Princess and South Korea), the number of completely asymptomatic cases is less than 50%. Using the most optimistic data, I personally have less than a 0.5% chance of dying assuming I get all the medical care I need.
And that's the problem. This virus hospitalizes about 20% of identified cases. They require some supplemental oxygen, and maybe an IV. With good care, probably less than 1% die.
So there are really two key fatality rates:
1. How many people die if they get all the care they need, and
2. How many people die if 30+% of the population catches this at the same time?
Even if we're overestimating (1) by a factor of 10, that's still enough to make (2) catastrophic. What happened in Wuhan and Lombardy can happen here, and there's absolutely no reason that it couldn't get 10x worse. Even if we're overestimating the disease.
So let's start testing aggressively for antibodies. Until we get that number, I'm all for extreme caution.
it doesn't matter what the IFR (infection fatality rate, usually impossible to know exactly) is. what really matters is how many people will end up in hospitals at the same time in absolute terms. after the first wave passes, it is important to know how many people are now immune to plan lockdowns and health care capacity for the second and future waves - but this again isn't related to fatality rates, but hospitalization rates and true R0 (maybe there were much more people infected who didn't even know they had it than we think).
We all assume that this is so (and I hope it is), but this is not completely clear. It might be that for covid-19, help from the medical system doesn't matter much, given our current state of medical knowledge and care. And iatrogenic harm cannot be discounted. Certainly we cannot ethically run the experiment.
As a single anecdatum, there was a (BBC?) piece on a hospital in the worst Italian region, showing a number of people on ventilators. The narrator explained that the staff was demoralized by the fact that everyone they had put on a ventilator died anyway.
Edit: CCP comrades downvoting this, here's how "zero new cases" is being achieved in your country: https://www.reddit.com/r/PublicFreakout/comments/fosodi/covi...
That said, the no new cases sure does sound like bullshit.
As does 50-60 cases a day per 1.5 billion people when your test accuracy is pretty much a random number generator (as Spain has unfortunately found out when they received Chinese tests). I'd have a much easier time believing 5-7K new cases per day. For China that'd be the same as 1K per day for the US, which for us would be not an issue at all, if it kept steady.
And also, we should hope that proper triage will mean that the non-coronavirus emergencies that matter will indeed be allocated a bed, etc.
I have little doubt that this will be awful, but the counterfactual scenarios are difficult to evaluate.
If the medical system starts to fail it's not just COVID-19 you get to worry about, if the incapacity rate is high enough you get to start worrying about things like: how long will it take for my electricity to get fixed, how long will my heat be out for, where's the food for the local supermarket, when will my pharmacy get my prescriptions in stock, etc. Second order effects start getting noticeable quick in a society as connected and just in timed as ours.
Most first world countries have hardened, resilient, and distributed infrastructure that takes little resources to maintain. The more your economy revolves around non-essential goods and services, the more headroom there is for dealing with failure.
The second order effects of breakdowns in supply chain are real. But they are most effectively dealt with by JIT and connectivity. The US is changing significant parts of it’s economy to deal with this shock: and it looks to be effectively redeploying resources without any leadership at the top.
Look historically at whole countries hit by war or disaster, and see how they managed.
Imagine if tomorrow all medicines are unavailable, and all hospitals destroyed? Many people would die, but the vast majority of the economy would keep running and the vast majority of citizens would remain healthy. Sure, we all could have stocked up on some resources better, but there are heavy costs associated with over stocking and over planning.
Based on my research, it actually seems like fatality rate will NOT skyrocket if our medical system is overwhelmed like in Italy.
Here is an actual study that looks at fatality rates in Italy vs China after stratifying by age group: https://jamanetwork.com/journals/jama/fullarticle/2763667
The most eye-opening graphic is the first chart. To summarize the article:
1) Italy's fatality rate for patients younger than 60 is actually better than China
2) Italy's fatality rate for patients older than 60 is less than double that of China
3) Italy may have a higher rate of pre-existing conditions which makes fatality rate look worse than it is
4) Italy is only testing symptomatic patients which makes fatality rate look worse than it is
All of these results were surprising to me. The information being spread by the mainstream media suggests that if hospital systems get overwhelmed like they have in Italy, fatality rate will skyrocket.
In reality, it seems like we're dealing with a classic case of Simpson's paradox (https://en.wikipedia.org/wiki/Simpson%27s_paradox). If you stratify data by age group, you find that even if hospital systems are overwhelmed (to the same degree as Italy), fatality rates will, at most, double.
The amount of people that need medical care caused by this virus should also be taken into account. And that rate is high.
While not a direct fatality of the viruses, these cases should be estimated to see its total mortality impact.
3,800 patients, or 12% of case in NY were hospitalized as of today. 888 require intensive care. [0]
You're right about approx 1% fatality rate though. And this is NYC the epicenter of the pandemic in the US.
[0] - https://www.nytimes.com/2020/03/25/nyregion/coronavirus-new-...
There is no Antibody/serological/immunity test yet. Biologically it is harder even that a vaccine... Test WILL be intensive/invasive (think blood, not mucus), much more so than a simple test for presence/absence of a particular virus.
https://www.reuters.com/article/us-health-coronavirus-immune...
Important though
Seems to miss any mention that people who get sick - but don't die - seem to be having pretty severe (sometimes permanent looking) damage.
Aka, they're only counting "deaths", when they should also be including other very serious negative consequences too. :(
I'm just pointing out that some people over the years have presented a very cynical view of US drug company leadership, and the way they influence things for their personal benefit regardless of human cost. :(
eg: Purdue Pharma (apparently owned by the "Sackler" family)"
https://en.wikipedia.org/wiki/2020_coronavirus_pandemic_in_I...
It's a different, though related virus, but there is some reason to believe that at least some victims will have lasting effects.
No, it is NOT pure speculation. We have plenty of CT scans showing permanent lung scaring in thousands of patients.
If it's true, link it. We have 20,000 dead it should not be hard to find.
The BEST I've seen so far is a study of 12 people
https://www.businessinsider.com.au/coronavirus-recovery-dama...
We are making the claim Covid-19 has a really low mortality, much lower than the 1% claimed. But even at 1% your link in not relevant.
So the question stands to the downvoters, one fucking link?
Are you all Russian trolls? Why do you think covid-19 does permanent damage. Where is this coming from? I've linked the only study I've seen.
People like the guy you're replying to are angry because their comfortable worldview is being challenged and hard.
If there are assumptions about Covid-19 damages I want a link.
The Russians trolls clearly know they can do more damage to the USA these coming weeks that normally would take years. They clearly will be here in some capacity.
This fear of damages to the body over death is clearly being used as a tool by multiple parties. I want them accountable. Don't care if it's the WHO or Russian Tolls or crazies. If they are right they will have links.
(And I addressed the above to downvoters not you sorry, people with comment history I trust)
I'm having mild symptoms, but it's been weeks, fevers come and go, and honestly it's like nothing I've ever had before.
And this is also another huge suspicious issue, for a killer disease it's not crippling many people as far as I've seen. Just crappy studies of 12 people.
If you take a look at Diamond Princess.
~ 4000 on board
* 712 cases
* 10 deaths
Gives a case fatality rate of 1.4%.
The average age onboard was ~60 years (don't know if it includes staff) which is definitely higher than average. However, the two numbers (1.4% and 0.06%) sounds very far apart.
So, no, 1.4% and 0.06% sound about right given the age difference.
Age | US Pop | CFR | Est Deaths if
| | | 100% infected
--------+--------+-------+--------------
0 - 9 | 40.01M | 0.0% | 0.00M
10 - 19 | 41.97M | 0.2% | 0.08M
20 - 29 | 45.43M | 0.2% | 0.09M
30 - 39 | 43.63M | 0.2% | 0.09M
40 - 49 | 40.46M | 0.4% | 0.16M
50 - 59 | 42.83M | 1.3% | 0.56M
60 - 69 | 37.41M | 3.6% | 1.35M
70 - 79 | 22.66M | 8.0% | 1.81M
80+ | 12.68M | 14.8% | 1.88M
--------+--------+-------+--------------
Total |327.08M | 1.8% | 6.02M
[1] https://ourworldindata.org/coronavirus [2] https://www.statista.com/statistics/241488/population-of-the...
= .3%
If it can't really kill anyone under 70 or 80 (Seems kinda right) you might have to look at the demographics deeper.
You need more numbers.
It will kill many more people under 60 where we run out of hospital beds and ventilators.
However it bothers me that there are no solid numbers at this point that properly controls for age and prior conditions.
It really should be possible now and we shouldn’t be discussing whether this is real or overblown.
In a confined space that everyone is stuck in for an extended period of time, people are continuously getting exposed and re-exposed to the virus, and those people are getting infected and continuously re-exposing everyone else. The ship eventually becomes a frothing virus stew.
Maybe one exposure ends up with a bit in your mouth from a nearby cough, and then another from your hand touching a surface and then wiping your nose or eyes with the back of your hand, etc., and then those keep happening over and over. Without proper ventilation, everyone will be getting exposed from breathing the same air in certain rooms, too. The frequent exposures give it more numbers that need to be fended off and more footholds to gain in different internal areas.
Then combine that with the high average age, lack of adequate medical care on the ship, general ignorance over the virus at that point, maybe re-use of unwashed clothes or hats.
Extrapolating from an environment like a ship or a hospital should be done carefully.
Meanwhile, https://www.worldometers.info/coronavirus/ shows Deaths: 19,603
I'm not an epidemiologist but how can we barely at the mid-point of the spread and say things like the above, while the hard facts already show we're at 20k deaths?
While I agree the actual rates are hard to know until we have robust antibody assay in larger populations, it seems a bit hand-wavy to say "oh it's just not that bad, here look at these random samples we looked at it and it'll be ok."
https://joindiaspora.com/posts/6b0f5a00507f013830be002590d8e...
The specific number based on a projection from 11 days' data on March 6 would exceed all deaths in all wars in all US history.
The deaths projection specifically has run high (by about six days), though case counts remain surprisingly consistent with reported data. There are plenty of reasons to be cautious about the data and trends. And there's no necessary reason this possibility must come to pass.
But if the US continues down the path it's chosen to date, that is what is in store.
I think the answer lies in between, which is the reason for the unraveling of this panic.
It's probably a poor estimate, but if we could extrapolate from the US population to the world population, 1.28 million died from the flu that year (61e3*7.8e9/372e6).
[0] https://www.cdc.gov/flu/about/burden/index.html
[1] https://www.worldometers.info/coronavirus/country/us/
EDIT: I found another link with a better estimate for world deaths from the flu. According to the World Health Organization, between 290,000 and 650,000 die from the flu each year [2]. So my estimate above was off by a factor of two.
[2] https://www.who.int/news-room/fact-sheets/detail/influenza-(...
0 785
3 1570
6 3140
9 6280
12 12560
15 25120
18 50240
21 100480We should know in 21 days if your prediction is right.
1. The virus begins to exhaust the available fuel (i.e., something approaching 60-70% of the population gets sick), or
2. Strict social distancing measures are implemented.
Option 1 means that ~200 million Americans contract the virus, ~20 million need hospitalization, and ~2 million die. At that level of hospitalization, though, the healthcare system will collapse, and the death rate (not only for coronavirus patients) may well go up significantly.
That's the logistic curve you'll get if you leave things to run their course.
Do you have a strong prediction for the future with a date and count? I'm not sure what qualifies as "strict", and I'm only interested in counting deaths.
The outbreak is growing exponentially at this point, and basic epidemiological models say that it will continue to do so until a significant fraction of the population becomes infected. COVID-19 is estimated to have an R_0 of around 2.4, meaning that if nothing is done to lower transmission, the epidemic will only begin to recede once approximately 1-1/2.4 = 0.58 (58%) of the population is infected.
The current doubling time for the epidemic is 3 days in the US, meaning that the US will reach the saturation level of ~58% of the population infected sometime this May.
In the face of the cold logic of exponential growth in epidemiology, you can't try to reassure people by pointing to the current numbers of infections. You have to look at R_0 (~2.4), the level of preexisting immunity (0%), and the mortality rate (~1%), and act accordingly. I don't want to see 2 million Americans dead by the end of May. Pointing out that only a few hundred have died so far and asking me to ignore basic epidemiology doesn't reassure me.
Are you predicting 2 million Americans dead from COVID-19 by the end of May?
If you feel confident in your predictions, make a strong claim. If you're less confident, put some error margins around your claim. If you need to put some caveats (i.e. "strict"), be specific. If you're uncertain, it would be more intellectually honest if you indicated that when you post.
You're going around suggesting that maybe we can ignore basic epidemiology. There is a strong push right now in the United States to play down the epidemic and get people back to work.
The best information available right now indicates an R_0 of between 2 and 3. With 0% preexisting immunity, that means that around 60% of the population will be infected, unless strict social distancing measures are implemented. With the observed doubling time of about 3 days, the epidemic will burn through the population by early May. The mortality rate is around 1%, based on the available information, meaning that about 2 million Americans will die.
This can be averted, but not by pretending that epidemiology is potentially all wrong. It can be averted by things like the shelter-in-place orders made by certain governors.
I'm currently much more concerned about what will happen economically and politically because of this thing. There are real and awful consequences to all of this panic that have nothing to do with dying from the virus.
Several states have already issued shelter in place orders. From what I can see, my state (which has not) is taking it very seriously regardless. So you get to keep doing your fear mongering predictions of 2 million deaths, and then later you get to say, "well yeah, that didn't happen because we sheltered in place".
You responded to DanBC's comment, which extrapolated the exponential by a few weeks, by suggesting that he might be wrong and that somehow a logistic curve will set in.
I responded to you by pointing out that the only things that will break exponential growth are herd immunity or strict social distancing measures.
> I'm currently much more concerned about what will happen economically and politically because of this thing.
That's exactly what I suspected. I think your line of thinking, which is being pushed by a lot of people now (beginning with the President of the United States) is dangerous. You're asking us to roll the dice and hope that either the laws of epidemiology will cease to hold, or that the mortality will turn out to be much lower than is currently estimated.
The available information indicates that the death toll in the US from business as usual will be around 2 million people by early summer. That merits a strenuous response, even if it does damage the economy.
> There are real and awful consequences to all of this panic that have nothing to do with dying from the virus.
We're at the beginning of a pandemic, with 0% immunity, 60% of the population expected to become infected within several weeks, and around a 1% mortality rate. A bit of panic is justified. Better than panic, however, are strong measures implemented quickly. Luckily some states have taken those measures now.
> So you get to keep doing your fear mongering predictions of 2 million deaths, and then later you get to say, "well yeah, that didn't happen because we sheltered in place".
You simply can't accept the number of 2 million deaths because it's outside of your experience. It is what epidemiology predicts will happen without measures like sheltering in place. How do you think we'll avoid 2 million deaths otherwise? This virus is going to infect ~60% of the population in short order if left unchecked. Do you think the virions will suddenly decide that they don't feel like infecting anyone else?
I really do hope that I get to say in the end that "that didn't happen because we sheltered in place." That's what people are already able to say in China, because they sheltered in place and stopped the epidemic.
You're asking us to roll the dice and hope that the best information about the R_0 and mortality rate of the virus is incorrect. I'm looking at what happened in Northern Italy, where the hospital system was overwhelmed in two weeks, and I'm not willing to do that.
The rules apply even though things are in crisis right now. Indeed, they apply even more because things are in crisis right now. Please review https://news.ycombinator.com/newsguidelines.html and stick to the guidelines from now on.
"Do you think the virions will suddenly decide that they don't feel like infecting anyone else?"
Note how I said "repeatedly". There are other swipes elsewhere in the thread, and a pattern of doing that in the past that we've had to warn about.
The cumulative deaths are now doubling at a slower rate (something like 5.5 days?), partly because stronger social distancing measures have been put in place.
But also, let's see what the CDC says next week, because those figures are laggier but more accurate.
15 days ago was just after New York and California implemented shelter-in-place orders, so we'd expect to see levelling in the numbers of daily new infections starting about two weeks after (now), with levelling in daily new deaths about 10 days after that (about 20th April).
Here's what other people are predicting - a peak on April 12th with over 2000 deaths in one day, and total cumulative deaths just over 23000 . https://covid19.healthdata.org/united-states-of-america
Why are you stating this now, and not when you wrote the 100,000 number above? If that wasn't a prediction, what was it?
What were you doing when you said "there are 785 deaths so far in the US from corona virus" and compared that to a figure of 61k flu deaths?
(Leaving aside the fact that you're counting covid and flu deaths differently in your comparison -- over counting flu and under counting covid)
These predictions are always shown to be exaggerated after the fact, and then the doomsayer gets to say something like, "Yeah, if it weren't for all of us telling you" or "You're ungrateful for disregarding all the hard work of everyone who came together to blah blah". There's no accountability for, or end to, the bullshit.
Moreover, while it's a morbid task, rational adults should be able to ask whether the costs we've paid dealing with this corona virus are worth it. At 758 deaths, and even 21 days later at 26,300 deaths, we've wrecked the lives of millions of people in the US. It's hard to find stats, but tens of millions of people are now unemployed, and many of their lives suck because of it. Doing arithmetic on suffering and death is distasteful, but there is some point where you wouldn't ruin N people's lives to save just 1 person from dying. Particularly when some of those N people are going to commit suicide in response etc... I won't be making any predictions, but other really bad things can happen when you've got a huge number of people who can't afford food or rent.
Two years ago, approximately 61,000 people died from the flu in the US and we didn't do anything more than make flu shots more available. As for over vs under counting, I hope you're not doing that pedantic "pneumonia isn't the flu" thing. The point is that 150,000 people around the world die every DAY from something, and 10,000 - 100,000 people in the US die every year from something with symptoms close to corona virus.
If you're going to mention flu deaths please at least count covid-19 deaths using the same method.
> people died from the flu in the US and we didn't do anything more than make flu shots more available.
We have internationally coordinated campaigns of surveillance and data gathering. We have rapid vaccine development. We have global, regional, and local flu strategies. We have pandemic preparedness programmes. We put in place public health measures -- you may not have seen those but they're there.
> Moreover, while it's a morbid task, rational adults should be able to ask whether the costs we've paid dealing with this corona virus are worth it. At 758 deaths, and even 21 days later at 26,300 deaths
Do you know that you're undercounting covid-19 deaths? The number you quote is deaths in US hospitals. Many people die with covid-19 outside hospitals, mostly in care homes. This is partly because hospitals in some places are overwhelmed and they're triaging elderly frail people onto palliative pathways. And yet we still have people like you denying the reality by quoting statistics that you do not understand: covid-19 is causing massive excess mortality.
I sadly had to pop his bubble by pointing out that value represented more than the sum of money in the entire world.
What drives exponential growth in an epidemic is the fact that each person infects some number of other people, on average. The more people are sick, the more people get infected each day. Growth is proportional to size, so you've got an exponential curve.
What stops exponential growth in an epidemic is when so many people are immune that the average sick person isn't able to infect more than one other person. At that point, the epidemic dies off exponentially.
You only reach the exponential decrease phase once a significant fraction of the population has immunity. That only happens once they've gotten sick and recovered.
Based on the transmissibility of COVID-19, it's expected to eventually infect 60% of the world population, unless a vaccine is developed before it does.
Your point was obvious, but if you meant to downplay point made by the person you were responding to, then you were wrong to do so. The exponential can indeed be extrapolated for many e-foldings, until the number of infected people is a significant fraction to the total population.
I'm not a huge fan of sarcasm, but I think zepolen was right to point out that blindly predicting exponential growth is ridiculous. The world is more complicated than some sophomore level differential equations would indicate.
> If you've got a real prediction, why don't you say it?
The doubling time of the number of infected people will continue to be about three days, until strenuous social distancing measures are implemented (or in the worst case, herd immunity is reached). The rate of newly detected cases will begin to bend downwards about 1-2 weeks after the introduction of such measures, due to the delay between infection and onset of symptoms.
> And you moved the goal posts from "deaths" to "infected".
The number of deaths is proportional to the number of people infected.
Proof: China's infection rate https://www.worldometers.info/coronavirus/country/china/
If R0 is in the range of 2-3 the interventions need to reduce transmission by 75% to 85% to get R0 values down to 0.5. At that rate exponential decay starts and the math works the same in the other direction.
The China or South Korea data isn’t especially comforting when you look at the scale of interventions required to bend the curve.
https://www.businessinsider.com/chinas-coronavirus-quarantin...
- Coronavirus disease 2019: the harms of exaggerated information and non‐evidence‐based measures[1]
- A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data[2]
There are two ways of getting the answer to the question: get relevant data to make reliable predictions (e.g. test for prevalence in the population) or run the experiment (which some governments seem prepared to do)
1- https://onlinelibrary.wiley.com/doi/abs/10.1111/eci.13222
2- https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-a...
He's right to point to the Diamond Princess Cruise ship for example, but his statistics are already out of date. As of yesterday, 8/712 infected had died; if you expand this to the Grand Princesss, it/s 10/800, which is at 1.1-1.25% CFR. This is a lot bigger than the statistics Ioannadis cites, and greater than the flu. There are some issues about how to extrapolate the cruise ship population to other populations, but that isn't straightforward (cruise ship passengers are probably older on average, but also might be healthier and/or wealthier). In any event, that number can only increase.
The other issue is that regardless of CFR in the population, the experiences of various locations points to the threat. That is, we can argue with Ioannadis all he wants about CFR and so forth, but in the end, if places are crumbling under the weight of hospital visits, it doesn't matter.
As someone else pointed out, fatality rates are only part of the picture, and in this case it's important. At some level, it's not the the fatalities per se that are the sole problem, it's the much greater percent of cases that require hospital care. It's not like CFR is a fixed statistic that applies everywhere all the time anyway; it will depend on resources and other things.
Finally, there's been a lot of discussion by Ioannadis and others about costs of restrictions and so forth but little data about that that's pointed to (even though it's available). For example, if you look into data about the Great Depression and Great Recession, it appears that overall there was a decrease in mortality that followed. Data on health effects of recessions is actually pretty clear: although there tends to be an increase in suicide and mental-health related deaths (e.g., drug overdoses), there's a much bigger decrease in deaths due to things like motor vehicle accidents, cardiovascular events, work-related deaths, and so forth. So this benefit-cost mortality analysis, as morbid as it is, doesn't necessarily play out in the way Ioannadis assumes.
The hyperconnected world is like a baby's brain which is more connected than an adult. As learning happens connections are culled [1]. To the baby every new piece of info is mesmerizing or frightening. The brain hasn't yet understood how to process things, what to filter out, what to focus on etc
With this new networked world/hive mind, reactions to every new event are much like the reactions of a Baby as it blunders about discovering a new world.
https://www.edge.org/conversation/alison_gopnik-a-separate-k...
We must also know the number of fatalities, and this will only be known after the epidemic once we calculate the over-mortality compared to a normal year. So the current numbers are a low estimate too.
In the end we can tell nothing about the fatality rate which is the ratio of those values.
Hence governments make a guess and act accordingly. Only those who test large scale and isolate only positive people act rationally.
If we'd let this thing run its course, wouldn't almost all corona patients have to be refused in hospitals and instead die at home? Normal "triage" doesn't cut it. You'd basically paralyze normal health care for months, which would come with an additional death toll. This is already happening even with the lockdowns.
Technically speaking, it will help, a little. But to actually measure up to the challenge you'd need to completely replace the entire thing with something vastly more capable.
Perhaps the article was arguing that even with covid-19 deaths plus "health care paralysis" deaths, it's still not worth shutting down the economy. In that case my point is moot.
Also, the other side of the equation is said to be perfect, but as was shown the other day, a lot of elderly people are left for dead in their homes or retirement villages and not reported. Also, I don't know if all deaths in hospitals are checked for coronavirus.
The second (https://healthpolicy.fsi.stanford.edu/people/jay_bhattachary...) focuses on "the constraints that vulnerable populations face in making decisions that affect their health status, as well as the effects of government policies and programs designed to benefit vulnerable populations. "
This is not the time for folks to hit the pages of the WSJ to start developing their viral infectious disease modeling muscles. Imagine if you asked a compiler writer to start developing a commercial OS.
These folks fundamentally do not understand viral growth models. Their article demonstrates their lack of understanding and they keep falling back to the normal models (pun intended) of epidemiology based on the usual statistical machinery. This is a common problem in this analysis. I made a more detailed comment further down that demonstrates the pitfalls of this thinking.
(edits: grammar)
Of course policy-makers should continuously evaluate whether or not shutting down a whole economy is worth it. Definitely we need better testing to understand the true number of infections.
But you can’t completely disregard current human responses when a) the condition can be fatal (game over, there’s no retry) and b) the virus can leave behind long-lasting damage.
In this case a purely quantitative argument based on fatality statistics feels myopic.
For sure agreed we need to find ways of gathering more and accurate information.
See dramatic ramp-up in testing, and where the number of tests administered is significant, a reasonable approximation of the infection rate can be established. So the recent blow-out in NY was actually a good thing and it does not reflect the true daily infection rate: they just didn't know how many cases they had. They still don't in fact, asymptomatic cases don't get tests, and those who had the virus and now have immunity (of which Cuomo's advisors suspect there's at least 100K) can't be tested without the new serological test currently under development.
The only reliable metric of severity remains the number of deaths, and until that gets into tens of thousands (i.e. exceeds that of flu), any panic is premature. We're not Italy. We're not Spain. The current level of response to this is unprecedented.
Completely agree.
"until that gets into tens of thousands (i.e. exceeds that of flu), any panic is premature"
I completely and totally disagree with this statement. By the time you get to >10s of thousands of deaths (e.g., about 20 days ahead of the actual infection rate) before you implement severe social distancing and Wuhan-style lockdown, you are set up for a guaranteed million deaths by the time it's over. 10's of thousands of deaths is too damn late. The time to act was a month ago. We are already chasing the dragon, and a number of miracles will need to happen to get this thing under control without 100,000 deaths.
Also, you're making decisions based on pure panic in the absence of reliable information, and you're making them country-wide even though the country is not uniformly affected. The reality on the ground is some parts of the country need to remain more operational than others for us to pull through this, and aiming for sub-flu levels of fatalities is an unrealistic goal, no matter how much fear mongering you see from the press. It's simply not going to happen, even if COVID19 is cured entirely, if for no other reason that we can't cure the flu. 60K people die every winter, nobody gives a shit. 1K people die of coronavirus - everyone loses their mind.
Now granted, it could get much worse very fast, but that's why we're ramping up testing so massively: to be able to offer a more adaptive, more localized response that doesn't shut everything down.
"You can't shut down the economy for more than a few weeks before supply chains fall apart and you start running out of food and necessities, not to mention anything more complicated. "
The California shut down has no affect on food and consumer necessity supply chains.
"Also, you're making decisions based on pure panic in the absence of reliable information"
This is not pure panic and there is information, and we know EXACTLY how to model in the uncertainties and confidence intervals surrounding best estimates of CFR, R0, etc. We have very good numbers from S. Korea. We have very good numbers from Diamond Princess and we can stratify based on age and prior health status and cross-check with Korea. WE are not in a blind panic. The press is not fear mongering. We are running out/ have run out of ventilators in NY. It has already TODAY gotten much worse very fast, and it should be no surprise to anyone who was paying attention to Italy two weeks ago. 1K deaths today, following an exponential growth pattern makes the numbers VERY scary in two weeks and that's assuming our healthcare system is not overwhelmed.
"but that's why we're ramping up testing so massively"
Are we? Not yet? And why not? Because Trump didn't want his numbers to look bad so he ordered the CDC to insist all tests were run through them. Now our country is scrambling to catch up. We will ramp up testing, I am sure. It is in the works, but we still haven't actually done it on the scale we really need. NY is getting sort of close, but not really. Even they have to at least double. EVERYWHERE else in the U.S. is laughably behind.
Citation needed. If it has "no effect" then we should just get rid of whatever they shut down, since it's clearly not necessary.
> We have very good numbers from Diamond Princess
We don't. All the people who died there were in their 70s and 80s and had pre-existing conditions as well. And it was a single digit number. One would have to be totally statistically illiterate to extrapolate that to a country the size of United States.
> 1K deaths today
Yes, the slope can get pretty scary in the middle of a sigmoid. But there weren't "1K deaths today", and they are seeing gradual reduction in the number of daily new cases. This is especially encouraging given that they're doing a ton of tests now.
> Are we?
Yes: https://www.politico.com/interactives/2020/coronavirus-testi.... With half a million tests given so far, best I can tell we have the highest testing throughput in the world, and it keeps on increasing. It's very impressive actually. I wish we didn't fumble it so badly out of the gate, NY could be much better now if they had more reliable info earlier.
What we need is serologic test to determine approximately how many people had asymptomatic COVID19. Some studies suggest there could be a ton of those, which would explain why Wuhan did not experience a second round of epidemic after lifting most of the restrictions and putting people back to work. Assuming, of course, we're not getting fake news from there.
That is pure nonsense. Do you suggest an economy that only includes what is "necessary?"
"> We have very good numbers from Diamond Princess
"We don't. All the people who died there were in their 70s and 80s and had pre-existing conditions as well. And it was a single digit number. One would have to be totally statistically illiterate to extrapolate that to a country the size of United States."
Straw man much? I see you elided my whole point about using the VERY GOOD and comprehensive S. Korea numbers and also using them to extract the signal from the Diamond Princess numbers? What is your agenda here? To just confuse everyone?
"gradual reduction in the number of daily new cases" What are you talking about? This is flat out wrong. We had 13K NEW cases TODAY, an all time high for the U.S. and we are certainly only getting started. The numbers have been rising every day for the past week.
"With half a million tests given so far, best I can tell we have the highest testing throughput in the world" That is again, JUST PLAIN WRONG. That's .5M tests since JANUARY. Now, if we continue increasing at the rate we did this week, then we will have adequate testing in place in about two weeks. I hope that happens, but it's doubtful. Every increase has met with a new bottleneck somewhere in the chain.
"which would explain why Wuhan did not experience a second round of epidemic after lifting most of the restrictions and putting people back to work"
Again, what nonsense is this? Wuhan is not lifting "most" of the restrictions until April 8. They loosened a number of restrictions a few DAYS ago. The virus has an incubation time of 14 days. Are you that mathematically illiterate to see the uselessness of your already incorrect "fact?"
I've misread, sorry. I thought you were talking about Italy, which is seeing a reduction in the number of daily cases. So will New York, a week or so from now.
> nonsense
You need to choose your words better if you expect me to discuss this with you. At the moment you seem too frightened to have a rational conversation.
Again, that is completely contrary to any reasonable interpretation of the data we have. In other words nonsense.
I'll check back in a week and see if you are capable of having a rational, fact-based, conversation.
most likely those more susceptible get it first
it is not clear what percentage of the population would either not get it or if they do get not get noticeably affected
Here's a tip: it still comes out to be 5-10x higher than the seasonal flu.
With an R_0 of 2.4, approximately 60% of the population can be expected to contract the virus within a few months, unless mitigation measures are taken.
So for Germany it's 10000 today, but 13000 tomorrow and so on. That's what exponential means.
Let's hope it slows down now with the social distancing measures.
- There are about 28k ICU beds in German
- Usual occupancy is 80%, expectation is that 50% can be reserved for Covid-19 patients by delaying elective medical actions
- This gives us 14k ICU beds to treat Covid-19 patients
- Estimate is that about 5% of all infected will require about 7 days of ICU treatment
- This tells us, we can at most have 2k new patients requiring ICU treatment per day
- Knowing that the 2k are only 5% of total cases, the number of infected the health care system can take care of is 40k/day.
Obviously this calculation will change with more precise information about the estimated numbers.
[1] https://www.youtube.com/watch?v=Fx11Y4xjDwA (sorry, German only)
Absent that, there’s not much point of the shutdown. I guess it buys time for medical professionals to retrain and gather supplies, which is crucial.
People are willing to believe and hope for things that were already disproven in other coutries. Facts won't matter at all.
The economic harm arguments assume that the economic harm of the pandemic is confined to government imposed restrictions. When going to the movies carries the same risk as the most extreme sports it is reasonable to expect that most people are going to self isolate.
Of course given the lag it is possible given the growth curve that people could blindly find themselves jumping off a metaphorical cliff. At the same time I’d suggest that the significant amount of trauma the survivors would face would have significant repercussions for the economy.
This doesn't mean that the quarantine is pointless. Even if covid19 is as "harmless" as influenza, having two influenza type diseases is still worse than only having one.
I think there's not enough data available to evaluate if these lockdowns are justified. Especially the cost of the lockdowns are hard to assess. We won't know until it's over.
But even if the mortality rate is vastly over-estimated, it will be hard for a leader to justify that their country has twice the mortality rate as the neighbor country (even this rate is low, and even it a lockdown was avoided).
Besides, most countries are imposing a lockdown anyway, so "do like everybody else" is a safe bet for our leaders. Plus, the population asks for it.
It is a common error for folks in set A to believe that they can understand models built by folks in set C. It is a subtle but serious error for folks in set B to believe they can understand models by set C e.g. John Katz of Yale who is a doctor focusing on diets and wrote a woefully misguided article in the NY Times.
Roughly, I find the confounding factors about models about the current epidemic seem to the following:
1- 1% of the population severely affected etc. (see: Diamond Princess). What they fail to appreciate is that the Diamond Princess was an enclosed environment and the outbreak was controlled and limited to the population on board. In the general population, a virus with an R0 of 3 will infect 50,000+ people in 10 infection steps.
2- What most people miss is the collision of the 1% severity with the capacity of the medical system. This second-order effect is something hard to understand. 1% of the US is 3.6 million. Even if 10% of those cases turn up at the ER in the same year and occupy the beds for 2 weeks each it will be a disaster.
3- #1 and #2 interacts with life as usual demands on the health care system – accidents, heart attacks, strokes, etc. to create third order effects – more deaths as there are no beds and no personnel to deal with them.
4- PPE running short and causing infections amongst medical personnel leading to their quarantine, hospitalization or death (Wuhan, Italy, Spain) decimates their ranks and accentuates the stress on the health care system……. 20% of lost capacity translates to some fraction more deaths and more stress on the rest of the medical population.
I am sure there are factors I have missed.
We have seen many contrarian viewpoints. None of these contrarian thinkers make any concrete suggestions (Ionnadis, Katz, Friedman, Gillespie, Hanson) except making ominous predictions to how wrong we are and we need more data. Meanwhile ER doctors say this is the worst they have ever seen and bodies keep piling up. These contrarian thinkers provide no simulations of how saving the economy will lead to lesser deaths. Just hand waving and finger jabbing.
The contrarians seem to think that the playbook to deal with the epidemic has been improvised. Or that this is something modelers are thinking up on the fly. These playbooks have existed for decades with very good understanding of the dynamics and were used during SARs, Ebola and heck even back during the Spanish Flu. What seems to be lacking was parameters specific to Covid-19. Its R0, CFR, co-morbidities, etc.
The reason Taiwan, Korea and Singapore were able to act so decisively and fast was that they just dusted of the SARs playbook and knew almost exactly what to do. Taiwan enforced the first measures in early Jan.
I will take the word of the viral epidemiologists over the contrarian's armchair speculations and continue to overreact. I recommend you should too.
In other words, many of those who died today would have died within the year. Many were already in and out of hospitals for serious diseases.
So, not only is the IFR wildly unknown without serology tests. The CFR itself is also severely biased.
Currently in the U.S. you cannot get tested unless you have severe, life-threatening illnesses.
The fact that the U.S. has such a low CFR speaks very well of healthcare in the U.S.... considering the majority of those tested have severe symptoms.
The exchanging deaths vs. economic health thing is a subjective policy position, which is different, and doesn't necessarily change even if the fatality rate is lower than reported. If it still appears to be more contagious and more fatal than the flu, I think most people will continue to support the lockdowns.
If? The amount of magical thinking required to still have some doubt about this is undefinable. Spain is converting ice-rinks into morgues to store the freaking bodies. NYC has run out of ventilators and the federal response is something like: NY could have had some at a good price in 2015.
Pessimist: there are more cases than expected
They both mean the same thing. And for now, it is pure speculation.
People like that truly disgust me. Wouldn't shake hand.
We will not put a dollar figure on human life.
We can have a public health strategy that is consistent with an economic one.
No one should be talking about social darwinism for the sake of the stock market."
The statistical value of human life is a figure somewhere between $6 - $10M used to weigh social and economic policy. Using it to write off a percentage of the population dying as the "cost of doing business" is standard practice in social and economic policymaking.
Non-existent testing infrastructure and late/poorly enforced lockdowns have me convinced the approach to covid-19 being taken by the states completely ignores our reality. We don't have China's population control, South Korea's testing infrastructure, or Italy's population density.
There's lots of nuanced strategies we could've taken. We settled on the simplest, most hysterical, least sustainable one.
We should aim for a policy that fits with our society and not try to imitate regimes where other courses of actions are more viable (e.g. China has very good population control and thus can afford lock-downs, whereas Western nations have more emphasis on personal and civic responsibility and may have better ways of dealing with it).
I would love to say I trust our government and the decision they made, but I'm afraid I don't.
Imo we should prioritize helping at-risk population and allowing them to quarantine if needed, and let each person assess their own risk (with help) w/r to that virus and act accordingly, while making sure they have access to the resources they need.
As a side note, this follows a trend I notice in society to treat people as sheep that cannot think for themselves and must be protected (which is warranted for some people in view of the latest hysteria).
This article is pointing out that the fatality rate may be rather low and that nationwide quarantining may not be that useful in saving lives, instead we may need to focus on supporting hospitals and the elderly in other ways than mass quarantining.
They make some good points:
>First, the test used to identify cases doesn’t catch people who were infected and recovered. Second, testing rates were woefully low for a long time and typically reserved for the severely ill. Together, these facts imply that the confirmed cases are likely orders of magnitude less than the true number of infections. Epidemiological modelers haven’t adequately adapted their estimates to account for these factors.
There is a lot of work being done here and there are already antibody tests being used in research so I'm expecting to see some better modeling soon.
They also make some less good points:
>An epidemic seed on Jan. 1 implies that by March 9 about six million people in the U.S. would have been infected. As of March 23, according to the Centers for Disease Control and Prevention, there were 499 Covid-19 deaths in the U.S. If our surmise of six million cases is accurate, that’s a mortality rate of 0.01%, assuming a two week lag between infection and death.
I get the feeling that everyone who came from Wuhan to the US was incredibly aware of the state of their health and were ready to quarantine at the slightest snivel. At least the ones I knew were. It could very well be that the Jan. 15 case was the first or very close to the first to show symptoms. Also, the doubling rate of every 3 days is less accurate for the first several days of the disease which probably messes up this estimate even more. What I guess I'm getting at is that this is a messy back of the envelope calculation that is less useful than it is foolish.
>First, the test used to identify cases doesn’t catch people who were infected and recovered. Second, testing rates were woefully low for a long time and typically reserved for the severely ill. Together, these facts imply that the confirmed cases are likely orders of magnitude less than the true number of infections. Epidemiological modelers haven’t adequately adapted their estimates to account for these factors."
Those are NOT good points. Those are lousy points. The article strains credulity at every logical assumption required to reach their final guestimate. Each assumption is deliberately taken to arrive at a comically low number for CFR and a comically high number for how many have already been infected AND recovered. Moreover, the author is NOT even qualified enough to understand that he is not qualified to offer an informed opinion on this matter.
Are you a glass half empty or half full person?
The reasoning is based on estimating actual cases given confirmed cases. For example, in Italian town Vò, the entire population was tested to find a prevalence rate of 2.7%. Apply this to the whole province to estimate actual cases and then divide that by confirmed deaths. So assuming that unconfirmed cases mostly recovered without an event, the actual fatality rate goes down to 0.06%.
Arguably, the author doesn't have a lot of other strong data to back this up. Also, this would imply that a large part of the infected population simply recovered without needing to possibly sick treatment.
If this is true, however, it would mean we just had a 2 trillion dollar party :).
[1] there is speculation that greater exposure to the virus may make some people sicker
For reference, a 70 yo has a 1.9% mortality. A 80 yo has 6% mortality rate according to actuarial tables.
People and health professionals have pointed that out from the start and it should be hard to find anyone who doesn't know that by now.
> A universal quarantine may not be worth the costs it imposes on the economy, community and individual mental and physical health.
Sure, you can use expected utility theory to calculate that. What result you get depends a lot on the monetary value you attribute to a human live, or, if you think that's better, on the value of quality-adjusted life years (QUALYs).
If you're particularly sociopathic you can also reverse-engineer the simple models used in order to find the value judgments that will give you the decision result you want. Just tweak the values for QUALYs, lower the estimates here and there, and the result is that it's better to save the economy (or vice versa, depending on what you want). Call me cynic, but I'm sure plenty of people around the world are doing that right now.
It's not like there's been a shortage of folks who happily multiply the CFR figures by 60, 70, 80% of the entire population and present this as the inevitable death toll of, generally, the actions of the political party they oppose.
It is a factor though, and not a minor one. You test widely enough and you can track, trace, and quarantine your way to containment or at least massively flatten the curve and thereby avoid overwhelming the medical system and spiking CFR. That is exactly what S. Korea did. You didn't link to the article, so it's hard to tell if you are just a troll or not. If the article stated that as a factor, it is not wrong at all. IF the article stated that as THE reason, well it's pretty misleading. Did the media mess up here? We shall never know I guess unless you can produce a link.
My point was a bit different, that from the perspective of multi-attribute decision theory with standard additive models the trade-offs between various attributes only need to be consistent (~not violate axioms of additive models, if the model is supposed to be additive in the first place), but they can be based on any kind of value judgments.
Even if you use standard measures in health care like QUALYs and weigh this attribute in a way that was previously accepted, the attribute would traditionally only be evaluated within the normal health care budget, which doesn't apply here. It is not usually aggregated together with completely different attributes like economic costs over time, number of unemployed people over time, increase of homelessness, overall reduction of the quality of life for healthy people, and so on.
I fear that in the coming months people will pick whatever weights suit their agenda and come up with various models of "costs" here and there, instead of arguing clearly for their position and the underlying value judgments. This is not a new problem, of course, and it is further aggravated by social pressures and institutional habits/self-interest.
Moderators sometimes turn off flags when an article is good and the discussion seems able to be substantive (enough). That's arguably the case here, so we'll try overriding the flags. (This is not an endorsement of the article.)
In the whole world, there are only two places that can reasonably claim to be taking a scientific approach to those aspects of this virus at this point: Iceland and the Faeroe Islands. Both have tested about 3% of their populations (that’s about 6X better than South Korea, 12X better than Italy, and 100X better than the US). Iceland is trying to test a representative population, although their testing is still skewed towards ill and exposed people. The Danes of the Faeroe Islands are tracing almost 100% of cases.
We should still do testing of sick people to support healthcare decisions as we are today. We should do more testing of exposed people for containment like in South Korea and Germany. But that doesn't help much with the broader governmental and societal decisions we need to make. We must do more. Embrace the "and." No more "or."
Since everything is moving so fast, just keep it bold and simple: Race For COVID-19 Truth -- Test 30 Million People In 30 Days.
You may also not be surprised to know they're owned by News Corp.
I'm soon to be a non-subscriber, but I might have kept my subscription longer if the WSJ app had a setting to suppress the opinion section. It's not even interesting, and I'm far from a doctrinaire left-liberal.
Otherwise, you're in danger of just feeding your confirmation bias.
Can anyone or a mod enlighten me?
dang?
/s
as I see it this is BS propaganda from WSJ to normalize whats coming next, push to open businesses.
the main problem right now is that there is a lot of fear and uncertainty in everybodys mind partly because of Italy & spain. and as cases mount that fear will only harden so regardless of whether there is a lock-down or not people wont go out an shop until a sizable portion of population is infected and recovered OR we determine with american data that CFR is indeed quite low (doesn't look like from hospital testing data & they wont do widespread tests).
but regardless I do agree bailout package is not gonna be enough.
This is different to something like 2008 where people no longer trusted financial and other institutions.
This is more like pre-1970's recessions where the central banks squeeze the money supply for a couple of months. Or WWII mobilization. In each of those cases the economy came roaring right back due to pent up demand.
We will get the worst of both worlds if our response to this is to either a) let it 'run it's course' or b) try to make the finance sector whole. The US is looking at both those options and the result will be catastrophic of they come to pass.
Unfortunately, this too has become a partisan political issue with dems for shutdown/quarantine, and republicans for opening up.
If it does turn out to be devastating then those responsible won't face any consequences. When confronted with their weak response to the virus in any elections in the future the likes of Boris and Trump will counter that you're politicising the deaths of [thousands, hundreds of thousands, millions?] of citizens blah blah.
An early end of self-isolation would be an unmitigated political disaster for the Republican party. These are statistics so simple that even politicians should be able to understand them.
Edit: I would also like to add that it is very early days for covid-19. A significant percentage of the population will be infected at some point before November. How high of a percentage is inversely related to how long people self-isolate. The states that are barely hit at all today will the "worst hit states" tomorrow.
1. https://www.people-press.org/2018/03/20/1-trends-in-party-af...
2. https://statmodeling.stat.columbia.edu/2020/03/07/coronaviru...
3. https://en.m.wikipedia.org/wiki/Voter_turnout_in_the_United_...
This is a moderation issue because such flamewars aren't only off topic, but destroy the intellectual curiosity that HN exists for.
But asking for such a study is far away from the people asking to just let a few % of the population die because "it's for the economy".
[1] https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-a...