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).
I've looked at your other posts, and you seem like a reasonable enough person. I don't have a horrible beef with you, but in general I really despise it when people do the chicken-little doom and gloom thing and project gazillions of deaths on every latest horrible thing. Yes, awful things happen, but it's like a disgusting yearly sport for arm chair experts to exercise their righteous indignation at all of us while preaching how we're all going to die from Africanized bees, Zika, H1N1, whatever. Usually it's mixed with some other two-faced political agenda. You didn't really do that, but maybe you'll forgive me for thinking your exercise in exponential growth fell into the same category.
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.
The only things that will bend the curve downward are:
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.
My post above just stated the facts as they are today or a couple years ago, with citations and uncertainty ranges.
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.
Your post above compared deaths so far due to COVID-19 to yearly deaths due to influenza. That's a very silly comparison, because we're only in the very early stages of the outbreak.
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.
No one asked you to ignore anything, and my goal is certainly not to reassure you. Very silly or not, I'm not making predictions - I'm just keeping track of a few of you who are.
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.
This isn't some little game. This is a serious situation.
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 didn't suggest ignoring anything. Please quit putting words into my mouth, it's dishonest.
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.
We've had to ask you more than once before not to post flamewar comments to HN and not to cross into personal attack. You've done those things repeatedly in this thread. That's not cool.
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.
What in this very even-handed and factual post could possibly be construed as a “flame”? Just because this commenter is engaging with an interlocutor who obviously is flaming doesn’t mean that they themselves are. That’s just the all-too-familiar “both sides” fallacy.
Please don't post in the flamewar style to HN, regardless of how wrong or annoying another commenter is, or you feel they are. It degrades this place and makes for boring reading, especially when two users get locked in a tit-for-tat spat, which inevitably descends as it continues.
> Is an extra 2 million US deaths in 2 months within your experience?!?
No, but extremely well grounded and understood epidemiological models say that that is the likely result. I accept that things far outside my experience can happen, especially when there are strong scientific grounds for believing so. You don't want to damage the economy with lockdowns, so you're dismissing basic epidemiology. You earlier accused me of intellectual dishonesty, but your basic argumentation here is driven by what you want (not to damage the economy), rather than what the facts indicate (that an unchecked COVID-19 epidemic will eventually infect about 60% of the population and kill about 1% of those infected).
The intellectually honest argument for you to make would be this: "If the choice is between locking down the country for a month and damaging the economy, or letting the epidemic run its course and kill 2 million Americans, then letting the epidemic run its course is preferable." But you're not making that argument openly, and are instead trying to dismiss epidemiology.
> You say that as though it's "just some boomer's 401k"
I haven't said anything about "boomers," and in fact, I find it stupid when young people online dismiss "boomers." Baby boomers are going to be hit especially hard by this epidemic if it goes unchecked, and I think accepting damage to the economy is far preferable to letting the healthcare system collapse and large numbers of elderly people die.
> If the choice is between locking down the country for a month ...
There's something fishy there. If anything you're saying is true, one month can't possibly be enough. I'll skip the napkin math, but if your numbers are correct, it'd be 5 months if you could schedule people to get sick for available hospital beds perfectly.
(Ok, napkin math: I'm assuming everyone will eventually be exposed, using your 60% infected, 10% will need hospitals for 1 week, and 1 million hospital beds in the US. So pretty wide margins for error.)
> your basic argumentation here is driven by what you want
You and your straw men. I think you're inventing things for me to have said/wanted/believed because I haven't written much that you can actually argue against. Look at what I write, and quit making up things between the lines --- The only justification I've given for my position so far is that in my experience, predictions of doom have always been false. Maybe I'm wrong this time, and I guess we'll know in a 3-8 weeks.
> .. and are instead trying to dismiss epidemiology
No, I dismiss statistics made by dividing small and uncertain numbers by other small and uncertain numbers, and I always doubt extrapolation, even when it's based on good statistics.
> Baby boomers are going to be hit especially hard by this epidemic if it goes unchecked
And younger / middle aged people (without savings/investments) are going to be hit very hard by the economy collapsing.
> I always doubt extrapolation, even when it's based on good statistics.
The extrapolation is based on an extremely well tested and understood theory of epidemics. It's really as simple as: the number of people infected each day is proportional to the number of people who carry the virus, so the growth in infections is exponential. That's extremely solid logic and it's been proven time and again, so complaining about extrapolation doesn't make any sense.
> I think you're inventing things for me to have said/wanted/believed because I haven't written much that you can actually argue against.
I think your motivations are obvious. You keep dismissing the science of epidemiology, saying everything is uncertain, and you keep returning to the idea that the lockdowns will destroy the economy. A lot of people are arguing like you are.
> Maybe I'm wrong this time, and I guess we'll know in a 3-8 weeks.
At which point the US will have to go into lockdown anyways, only that the epidemic will be much worse. Better to do as doubleunplussed explains, and crush the epidemic now with a relatively short lockdown, and then partially reopen and try to contain the virus at a much lower level with widespread testing and contact tracing. China did this, and South Korea, Singapore and other countries did enough testing and contact tracing early on that they haven't had to go into lockdown.
You're right that social distancing until everyone gets infected is a bloody long time (I think you've underestimated it - other estimates are several years, though we'd probably have a vaccine by then).
The consensus on the best solution is to crush the virus with extreme social distancing, so that it doesn't infect most of the population, and so that we don't have to keep the economy shut for ages. This would only take on order weeks. Then we can open the economy up mostly, and keep hygeine standards high and do extensive testing and contact tracing to keep case numbers very low (low enough that we can still do effective tracing) until a vaccine arrives.
Sounds like a friend who in early 2018 told me that if Ethereum kept going up at even half the rate it had been for the next 3 years he would be retired!
I sadly had to pop his bubble by pointing out that value represented more than the sum of money in the entire world.
You have to ask yourself what drives exponential growth, and what eventually breaks the exponential.
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.
You made fun of the extrapolation of an exponential by extrapolating beyond the size of the human population.
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.
That is because China took drastic measures to control the spread of the virus. Many of these interventions were existing systems and cultural norms that were already in place due to the previous SARS epidemic. The US, for example, gave up on contact tracking fairly early while China had 1500 teams of at least 5 people doing contact tracking(and are still doing it)
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.
Yeah, that "can" happen. Other things "can" also happen... If you've got a real prediction, why don't you say it? And you moved the goal posts from "deaths" to "infected".
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.
It's not "blindly" predicting exponential growth. Epidemics grow exponentially, for very concrete and well understood reasons. There are a number of people making fun of or playing down that fact here, which is incredibly foolish. We're facing a serious epidemic, and we ignore basic epidemiology at our own peril.
> 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.
