The number of people who will die who had coronavirus or are strongly suspected of having it will be around 250,000.
The actual increased mortality won’t be that high; it’ll be between a third and half that.
That’s plenty bad enough.
Wow, less than 50k? Right now, the deaths are about doubling every 3 days. I bet it blows by 50k by the middle of April. Last year’s flu season killed an estimated 34k people. I have seen nothing to assume this will be easier than the flu.
Oh not an infection of 40% but a functioning as “resolved” rate - which puts the 24% of the population that are kids in that bucket.
Please appreciate the fact that this following the growth chart analogy also projects most of the impact occurring in the next 30 days, which will be beyond overwhelming, not spread over 4 to 6 months, like an influenza season is. The one plus side is that since we were later the most vulnerable in particular were to large degrees already socially distancing before governmental edicts. That may get us flattening sooner than Italy or Spain. And that our baseline is more socially isolated for more of our higher risk populations than Italy or Spain’s is. So maybe stretched out some more, one hopes.
Fauci just said that he expects in excess of 100k deaths, so I’m going with #4.
Cite: https://twitter.com/CNNSotu/status/1244275909944885248?s=19
I am not optimistic about the coming results of earlier “social distancing” in the US. People are barely taking it seriously now outside of the hotspots.
The mortality rate is a tricky statistic. The data seem to show that it’s possible to keep the mortality rate down to beween 0.5 and 1.0% in places where there is widespread testing, very early identification of positive cases, and immediate quarantine and treatment available. All three of these things have to be present in order to keep the fatality rate below 1%. Some countries are getting lucky – for now – in that they don’t have a lot of community spread disease. But once the disease gets into the community and once it gets piped into the transportation infrastructure and into the fabric of everyday living, this virus is off to the races and it’s going to score some touchdowns.
Right now, the virus is off to the races. We’re looking at a death toll probably around 5 to 10,000 within the next week or two, and there’s not much we can do to stop it at this point. After that, I think we’re going to see it slow a little. We’ll see the rate of infection slow, and we’ll hopefully see some cities catch up in treating the most extreme cases. Still, people are going to keep dying. Keep in mind that the case numbers now reflect infections from 7-14 days ago, which is when people started widespread social isolation - and only in some cities. I am hopeful that the number of fatalities will begin to slow. If we’re fortunate, the initial wave of deaths will stay somewhere between 20,000 to 40,000; if we’re not so fortunate, and if we keep getting in our own way, it could be over 50,000.
There’s a very good chance of a second wave, not necessarily because of cooler weather (as the media is foolishly reporting) but because of more social gatherings. Schools will probably reopen. We will probably gradually bring back commerce. Some sports leagues may try to have a shortened season. We will have Thanksgiving and Christmas. We will try to do all of these things with extreme vigilance, but there’s a good chance this won’t work. I think this smaller wave will not be as bad as this one, but it won’t be insignificant either. Our health systems will be better prepared, though this will certainly have an economic impact. If you look at the 1918 pandemic, the last wave (1919) was about 1/4 or 1/3 as bad as the 2nd wave (I’d argue that the fall wave of 1918 was really the first, as the Ft Riley outbreak was mostly contained initially).
So if we say that 50,000 people die in the first wave and another, say, 10,000 die in the second wave, that’s 60K, and I’d call that a success. If we continue to flub the responses, though, then we could have a lot more than that. I’m gonna say between 50 and 100K, with the possibility of going higher than 100K. I’d be somewhat surprised if it exceeded 200K, though not absolutely shocked.
One of the many many things that frustrates me is that I have no idea how seriously people around me are taking this . . .because I am taking it very seriously. Except for walks around the block, we’ve been out of the house I think 3 times in two weeks, for grocery store, doctor, take-out in some combination each time. So I have no freaking clue if I am the only one doing this, or if 90% of my neighbors are following the same protocols.
I make no predictions. I am unsettled by how wrong the people in the 538 article were; it suggests that the unknown unknowns are massive.
And your method of doing this is to poll random people on the Internet? You’d have equal efficacy scrying into your whiskey or receiving emanations from Urim and Thummim. Asking a poll of non-informed non-experts on this issue is like asking how many of the hundred pregnant women in your town are going to have miscarriages. This is beyond poor taste.
There is totally inadequate data regarding infectiousness, mortality, the percentage of the population which is asymptomatically infected, and the ultimate effectiveness of all of the efforts going on now to improvise protective equipment and respirators so that medical personnel can continue to treat severe and critical patients who may die without support. The only real guidance we have right now is an empirical look at the trends in other countries which had outbreaks proceeding the United States and the relative effectiveness of those measures to limit the spread and blunt the peak of COVID-19 cases, and by that measure, the outcome does not look good. The US is trending more like Italy and Spain rather than South Korea, and to say that this is not good is like saying that rain is a little wet.
All you can do at this point is take the best measures to protect yourself and your family based upon the best evidence that you have. The measures you describe are reasonable and at least reduce your risks far below the people who are standing around on a corner passing a vaping stick around (not kidding, I’ve actually seen this every day this week right outside my window) and despite all of the wealthy people suddenly buying out resort hotels or trying to travel to remote islands, staying home and limiting social contacts to the minimum necessary is really the best thing to do, because if Forrest Gump can’t outrun this virus, neither can a bunch of i-bankers and CEOs.
The FiveThirtyEight.com article doesn’t have a list of the ‘experts’ who made predictions so we can’t know what the basis for their estimates are or what political or other biases may be skewing the results, but it is pretty clear that the best answer is “We don’t know but possibly exceeding 2,000,000.” Of course, it will skew to the high side if measures are not put in place both to enforce social distancing and isolation, and to help support people whose livelihoods have suddenly been destroyed by those measures so they don’t go out and seek employment in ways that will obviate efforts at isolation. And until we have antibody testing to know who has been infected and what the distribution of people with acquired immunity looks like, we won’t know how safe it is to ease up on measures or let people who are demonstrated to be non-carriers return to work while still providing to support to those at risk.
Stranger
My point more had to do more with the idea that since the actual number of people who have been infected and recovered is a mystery, there could be lots of immune people walking around right now, whom we have no idea are immune.
If the number of cases in the US turned out to be, say, 10x the known number, that’s a lot of immune people, and if every diagnosed case represents 3 undiagnosed cases, AND we have a virus mutating very slowly, that will seriously slow down the rate of infection. The graph of infections may not be even on both sides. It may rise slowly with social distancing, but after it reaches a peak, it may drop precipitously, due not exactly to herd immunity, but something like, maybe “herd resistance,” as it becomes more an more difficult for the virus to find paths to travel.
Yes, that’s a “best case” scenario, but since we continually hear that there may be many, many more people infected now, or previously infected, I don’t think it’s impossible.
This exercise, distasteful or not, seems to be based on the assumption that a death will be definitely attributable to covid or not. But unless all people who get sick and die in the next months had been tested, you won’t know which.
Patients will present to hospital today in heart failure, with a COPD exacerbation, in a delirium, and with heart attacks. Some of them, although asymptomatic in other regards, will have covid. In fact, covid may have precipitated their acute problem.
A relevant NYT articleand one from JAMA Cardiology.
And, ftr DSeid, do you think those without health care who die without having interacted with the healthcare system will be adequately counted (or will have even been tested for the virus)?
The metric that will count - increase in all-cause mortality - won’t be available until it is ‘too late’.
So, the OP’s question is ill conceived (and may lead, inadvertently or otherwise, to cause people to underestimate the threat).
I believe it is informative to hear the thought processes of others, and to share my own, yes.
As to the group of, as you put it “experts”, who were part of the 538 bit - surely you can manage to click a link to their source and open the pdf?
Ya, you’re right, just “experts” with political biases.
Please be aware that NO ONE in this thread is encouraging letting up on any control measures at this time. In my strong opinion, the timing and nature of any let up of any sort has to be informed by actual facts, as we belatedly collect them, and as they occur. But this thread IS NOT about what should be done when and there are many other threads with that discussion contained. Take that discussion somewhere else and please do not hijack this one
The death rate is going to lag the case rate by 1 to 2 weeks. People dying this week were new cases 1 to 2 weeks ago.
Your point is dead on, though. If the death rate, adjusted in time to the diagnosis date, is climbing at the same rate or faster, then the increase in case rate isn’t just due to testing. I had included all that in my original reply but decided brevity was better.
KarlGauss, strongly disagree that an informed discussion about how bad the storm that is making landfall as we speak is going to be, and about how best to use the emerging information we have to understand that as best we can, given the lack of factual key inputs to run the usual models, is “ill conceived.”
No question that just like for influenza epidemics, like H1N1, full accounting will never occur, and initial estimates during the moment will be modified over future years. That includes eventually making estimates of those who died without healthcare system interaction, from other causes due to lack of services during a period of the system being overwhelmed, and those who died later because of care and diagnoses delayed by lockdowns. Still it was useful during H1N1’s emergence to try to understand what its landfall impacts were going to be.
If anyone gets the impression that the numbers need to be in the hundreds of thousands or more to be “serious” then they are not listening at all: 50 to 100K, (250 deaths/1,000,000 population) the range of a bad flu season, occurring as a surge in 1 month, which is the surge has looked like in Western countries ahead of us, is a nightmare scenario, in which a good many of them died because we failed to flatten the curve. IF that is a reasonable prediction, then slowing it down in communities not yet appreciably hit, so that it spreads over months and not weeks, requires taking action, well a week or two ago.
If you have some informed intelligent criticism of my thought that lacking other key information looking at the shape of the growth curves of death rates/million population (more likely a real number than “confirmed cases” which at this point is of near no value), or my interpretation of the shapes of those curves, I would be happy to hear it.
I agree with the sentiment. However, I think it needs to be discussed. The early estimates of it doubling every 6 days are not represented in the number dying. I’ve been looking at it from different angles. NY is the bulk of deaths in the US and their total deaths per million today (29-Mar) match Italy’s on 24-Mar. However, Italy’s numbers in the last week show less than a doubling every 3 days. NY is showing more. Obviously large/dense cities with high rates of international travel will be affected more. How well shelter-at-home works won’t be known for weeks. We broke even the last 2 days so fingers crossed.
I’ve been working with, well playing with different models, even the best one doesn’t look good.
I’m an Computer Science person, so powers of two factor into my thinking.
-
The naive, horrible thought: We just had a doubling to 2000 deaths after a couple of days. There’s time enough for maybe 10 doublings. That’s a factor of 1000. So, 2 million dead.
-
The better, much better thought. With currently 100+k known infected (and presumably a lot more unknown), you can’t really go 10 doublings and not get more people infected than total US population. So 10 doublings is clearly too much.
-
The not-so-naive semi-horrible thought. Let’s say we have maybe 2% of the population who might have it right now. Some 50+% are going to get it. Round that to 64 in binary terms. That’s 5 doublings. 5 doublings on top of 2000 dead already gives you 64k. Hey, getting into the ballpark of the UW guesstimate.
-
The really-less-naive-truly-horrible thought part 1: But there’s a lag between infection and dying. A lot of infected people are a couple weeks from dying. Conservatively, add two more doublings. 256k dead. Bye-bye UW guesstimate.
-
The really-less-naive-truly-horrible thought part 2: The current mortality rate is for health care systems that are under stress but not completely overwhelmed. That’s going to change. (We might be seeing this in NYC right now.) At best, that’s another doubling. But realistically, we’re looking at two doublings. 1 million dead. Probably more.
In the next few months.
The main hope is that the isolation stuff and what not helps part 5 quite a bit. But between the delay it buys and adding in the rest of the year, it’s a wash.
This poll doesn’t cover this side of things well.
Anyone thinking a flattening of the curve is close, just remember they only shut down Spring Break one week ago.
Florida also is a state with an older median age. Second biggest share of population 65+ in the country at 20.5%. (Only behind Maine.) Median age 42. (Four years more than for the United States as a whole.) It is going to get very bad there very fast I fear. No matter how locked down they are now.
the lag time between deaths and infected works both ways. As the number of deaths level off that’s an indication that the number of infected leveled off earlier.
I don’t put much stock in estimates of infected. Unless everybody is tested every day it’s an estimate based on those who have been tested.
If you ever played “lunar lander”, you know that waiting to long too fire your retros results in a very bad outcome; close to exponentially bad. That’s what we’ve done with COVID-19.
We’ll find out what the k is in a = e ^kt by fall.