We’re now past 63,500. Only about 5000 of those deaths were in March, so the U.S. has had more coronavirus deaths just this month alone than the entire American death toll for the Vietnam war.
What I saw today was even worse. That University of Minnesota expert’s assumption that at least half the country will contract it over the next eighteen months looks more and more likely based on the stupid way people are behaving. There was a thread here a few weeks ago in which someone was castigated for wanting to have an outdoor BYOB barbecue among neighbors, with everyone sitting fifteen feet part. That looks like incredibly conservative, super responsible behavior compared to what I’m seeing. :smack:
They now project a total of 72,433 deaths. By Wednesday, you ask. No, by August.
That model seems to be well off, and has been that way for quite some time. Death rates might decline a little but given that these relate to cases reported by up to three weeks ago I just cannot see it.Its off today by over 5000 and their projected total of 72k by August is not realistic as noted by RTfirefly
Seems to me that 90k by August is only a hope, and we have not yet seen the effects of protests, or incidents such as the funeral in NY where 1500 gathered to mark the event, nor the effects of the latest and future unlocks of states.
In the last 3 weeks the rate of reported infections has not declined - it is a steady progression, there would need to be a significant change to recovery rates for the number of deaths to decline - I doubt that the model has that sort of change built into it, and we do not yet know the full effect of the Remdesvir thing - it may be useful or it might just delay the inevitable.
-
Most of the models only function as projections under an assumption that current distancing measure are continued.
-
No question that after a good long period of performing well for the national picture at least the IHME model is now not doing so great. Thing though is that they had a way to lose. It is really hard to find any model with a decent track record. The linked graphic in the NYT piece informs some and here’s a CDC page with some recent past model projections. Pretty much they’ve all been of little predictive value and I’d love to see some head to head performance evaluations. The range of the forecasts looking even four weeks out, knowing all that is now known, is outrageous.
OK, but the IHME model has been updated since assorted governors had announced that their states were opening up this past Friday. If your model incorporates assumptions known to be false, sure, of course it’s going to be off the mark, but that doesn’t excuse it.
Tru dat, but having lost their way, they don’t seem to be interested in finding it again.
Maybe so, but the others were so widely cited that this is the first I’m aware of them. The IHME model was the one everyone knew about.
As I’ve noted elsewhere, one of the biggest issues is that they assume a normal distribution, and there is an abundance of evidence that that is not a good model for this pandemic. Do any of these “Daily New Death” charts look like normal distributions:
Italy
Spain
Belgium
Germany
China
US
I could do this all day long.
The Imperial College one on 3/16 was the first big news one that was widely cited, discussed here, and given how much you follow this and like the numbers I am honestly surprised you hadn’t heard of it. It was projecting over 500K dead in the U.K. and over two million in the United States … required was 5 months of heavy restrictions and then continued periods of intermittent restrictions for about two thirds of the time (believed to likely go on for at least 18 months). If 5 months of complete lock down was not possible they stated “that the alternative relatively short-term (3-month) mitigation policy option might reduce deaths seen in the epidemic by up to half” (i.e. still predicting a million deaths in the U.S., 250K in the U.K.). Original article - pdf. That changed to 7 to 20K in the U.K. by early April. Which given that it is currently 28K may have been a bit of an over correction. Revisions since. Others have gotten intermittent press.
My WAG is that IHME started to get more press because Fauci went with it, when they had adjusted up to a best case of 100 to 200K deaths in the U.S. early last month, and then stuck with them as they revised back down and down to lower than where they had started (now obviously revising back up again). But I think he went with it because it had been to that date the best performing of the lot.
IHME adjusts their estimate a wee bit:
It was well known back then that ‘they’ were saying potentially 2.2M people could die of the coronavirus here. I have no recollection of what source was attached to that figure. But for a couple weeks before I posted about it here, the only projection I’d seen anyone refer to was the IHME, and I’d seen it widely referred to.
Anyhow, it has been crazy low for weeks now, and they agree with that.
400,000
Yup. For anyone not tracking, they basically doubled their numbers over night.
It looks like they finally accepted a skew distribution to the epidemic rather than a nice normal distribution.
They still seem to think that hospital utilization and fatalities are in perfect phase, which is obviously wrong.
Well I reckon their revised number for August 4nd is also significantly wrong - its likely this will be realised by June 1st.
Add totals for June and July and you would imagine that approaching 200k is not unlikely especially given that the rate of infection is declining only very slowly and could quite easily rise during easing of restrictions.
That model has been consistently wrong, or maybe to be more charitable it is around one month to six weeks behind the reality.
Given that the predictive model has been so wrong I can’t but help wonder if there is a political reason.
Imagine the presidential campaign running against a background of 200k plus deaths with a very distant hope of an end in sight. Unlocking certainly runs that risk and the pressures of the economic impact will only increase the pressure to unlock further.
My guesstimate would be #5 from the OP; in the 300 to 500K range. The record so far seems to indicate that pessimism is likely to be the more accurate choice, here.
I’m sure a lot of Trump supporters think all the models that wildly over predicted infection rates were for political reasons too. None of the models have been very predictive and all have made major changes in their projections. Read the links DSeid provided above.
Maybe it’s not fair for me to jump in late, but I’ll go with 200K - 300K. There are about 200 days left this year. Figuring an average of 1,000 deaths a day, that puts the year end total around 300,000.
DSeid, you haven’t weighed in for a while. I never really understood exactly how your visualization of the trends in the death curve worked, but it seems like it was too optimistic. Is that fair?
For my part, it still looks like my prediction of right around 500K was too pessimistic–let’s hope so anyway.
OTOH, if this model (which didn’t get much national attention that I saw, surprisingly) released ten days ago is anywhere close to accurate, American deaths could be in the *millions *by next March (assuming Minnesota is not disproportionate to the rest of the country, and I don’t know why it would be):
They are projecting that more Minnesotans will die each day in July than have died to date! :eek: Hard to believe this could be remotely possible, but these are serious people making this prediction.
There is zero reason to put any faith whatsoever in a projection for next March.
I was watching the numbers from Europe back in March or so. I think Spain had the worst ratio among “closed” cases—at one point about 55% got over it and 45% died IIRC. Today that number among resolved cases is 13%. Why?
I’m guessing that initially it tore through the groups that were at risk. I also remember reading about doctors in Italy being killed by it—probably a lack of PPE. Since then I’ve heard that putting patients on their stomach instead of their back may help. Would CPAP machines have been better than ventilators? We’re learning as we go.
The OP asked how many would die from it. We’re going to see lots more infections, I’m sure, and from what Fauci et al say we can’t realistically expect a vaccine soon. It seems that we’re losing a lower percentage, at least, and with have time to study it and pool what we know with other countries, maybe there’s some hope.
At the time I had thought I was being a bit pessimistic … but clearly it was wrong, an extremely significant undercall, as we are past it already. I am at least relieved that curves have been flattened enough (and responses to make capacity strong enough) that systems have not been overwhelmed. My current thinking though is increasingly fearful of what will be in the late Fall, especially in regions that have seen relatively few cases to date.
As to the model predictions … there has been woefully little progress in pinning down the key inputs. No model has had a great track record, many have flipped around all over the place, and the only predictions that have not been way off have been for the very short term projections. I agree with CarnalK’s lack of faith.
But who knows? If it turns out that the common cold coronaviruses do impart some immunity/resistance, maybe any resurgence of covid-19 will be blunted by arriving during cold season.