Obviously, I'm talking about professional predictions. If for some reason you care about armchair predictions, it's going to be 10 to 1 in the "way worse" range for a random sample here.
Which accurate pre-peak predictions are you referring to?
I am talking about predictions from politicians not random people online. Professional predictions had huge error bars around various levels of response and tended to be surprisingly accurate IMO.
Hell the stock market still seems to think it’s no big deal.
Those underpinning the ‘flatten the curve’ projections and causing everything to shutdown for one. With exponential growth before a shutdown and flat growth during it.
There where plenty of different models based on different government responses, and they generally hold up well.
Suppression. Here the aim is to reduce the reproduction number (the average number of secondary cases each case generates), R, to below 1 and hence to reduce case numbers to low levels or (as for SARS or Ebola) eliminate human-to-human transmission. The main challenge of this approach is that NPIs (and drugs, if available) need to be maintained – at least intermittently - for as long as the virus is circulating in the human population, or until a vaccine becomes available. In the
3 case of COVID-19, it will be at least a 12-18 months before a vaccine is available . Furthermore, there
is no guarantee that initial vaccines will have high efficacy.
That’s the current UK strategy and they don’t have exponential growth.
Why wont you talk about the predicted numbers? It would be more useful than cheap shots. The paragraph you quoted is unrelated to the prediction.
p16: "In the most effective mitigation strategy examined, which leads to a single, relatively short epidemic (case isolation, household quarantine and social distancing of the elderly), the surge limits for both general ward and ICU beds would be exceeded by at least 8-fold under the more optimistic scenario for critical care requirements that we examined. In addition, even if all patients were able to be treated, we predict there would still be in the order of 250,000 deaths in GB, and 1.1-1.2 million in the US."
We achieved better results than the best case estimate for the US in that paper, and not by a little.
The prediction is not based on suppression but mitigation, thus we are outside the model.
But for now suppression is holding steady in the US and UK. With current rates projected out for 18 months being lower than the predicted mitigation strategy.
That said, there seems to be significant pressure to reopen, so we are likely to experience the mitigation strategy starting in a few months.
Suppression can limit spread to an arbitrarily small number depending on when it’s implemented. They correctly stated it’s capable of stopping exponential growth and that each mitigation or suppression strategy’s effectiveness depends on when it’s implemented. So, if we started a week earlier the daily death toll would be ~1/10th as large.
That prediction exactly matches observed outcomes. Do you expect them to predict our public policy and it’s date of implication?
There are no studies that I am aware of that predicted the exceptional outcome the US has had as far as total deaths. If there are, please link to them. The study that you think is suprisingly accurate, has a minimum number of 1.1M dead if we implemented "case isolation, household quarantine and social distancing of the elderly" in the US, which is a reasonable approximation of what we did.
Quarantine is harmful, and only useful if we’re preventing something worse. That’s probably the case, broadly — but the policies in other countries raise questions about US policies.
This is compounded by the arbitrary nature of closures, eg public construction and landscaping can continue but private can’t. People resent politicians arbitrarily (and self-servingly) picking winners and losers in the economy.
It’s further compounded by the fact that leaders haven’t announced a clear vision for when this ends. People complied because they cared and wanted to prevent deaths — but now we’re in a situation where the workforce is devastated and being hurt by arbitrary requirements, with no clear plan.
Naturally, people aren’t just going to sit there hurting — and if leaders don’t articulate a clear direction towards fixing things, people will find their own.
It’s also unclear whether the lockdown is accomplishing anything. We have enormous economic damage, lots of boredom, no dramatic decrease of cases in most places, an unimpressive amount of progress toward herd immunity, and no meaningful increase in testing, prevention, treatment ability, or clever policies from any level of US leadership.
In other words, what have we accomplished by locking everything down? Delaying large numbers of deaths by a few months?
Extending the lockdowns until we maybe have a vaccine in 2021 does not seem like a great idea.
What does that even mean ? It's slowly going down everywhere with serious lockdowns instead of going exponentially up. Anything less than exponential is a _dramatic_ decrease.
Exponential growth for a short period is seen in every epidemic, even small ones, because they all follow the same bell-curveish sort of shapes. Having observed that by itself doesn't mean anything - all epidemics grow slowly, then quickly, then slowly again as they approach their natural saturation point, then go into decline.
The idea that COVID-19 would grow exponentially starting from zero until a saturation of 100% was reached was only ever the outcome assumed by simulations, which have - as far as I can tell - a perfect track record of total failure. At least the sort of models created by universities in the modern era do.
To make this clear, the UW model driving much US policy response was updated several times in real time yet despite that managed to make wrong predictions for the exact same day it was updated.
And not just wrong, but reality fell far outside this models own uncertainty bounds too, so according to the modellers observed reality was not just unlikely but almost completely impossible.
The Imperial model that drove the UK ignored the possibility of hospital capacity increases entirely, and comes from a team that has (as far as I could find out) never successfully predicted an epidemic.
Real diseases don't play out like unending exponential growth until the whole world is infected - they go into decline far before that.
Quite why this is so is unknown. The dark secret of epidemiology is nobody really understands viral spread. Every model is wrong and nobody can explain why. Germ theory must have a major gap, so what is it? My guess is a combination of much higher asymptomatic spread than currently understood (asymptomatic infection not being a phenomenon that's really been investigated much as only academic modellers care about it), and perhaps some forms of natural immunity to viruses that aren't based on antibodies.
Whatever the cause, it'd be wrong to compare the observed outcomes vs the predictions from models that are known to be completely useless and say, there's been a decrease.
The answer is that the current US administration wants to draw this out until November when they can blame the economic toll of this disaster on the opposing party
We need this data now, but let’s also not forget this area needs regulation (that among other things will give teeth to violations of “but it’s all encrypted/anonymized!” claims)
This is particularly frustrating to witness in Australia, where we have all but eliminated covid-19, but there is no political will to eliminate it entirely. Had we closed hairdressers, building sites and takeaway restaurants a month ago, we would be looking at zeroes for daily infection.
While many places in the world cannot yet justify the costs of elimination, a country like Australia, bordered by water, could have reached the goal before fatigue set in. As it is, I am resigned to awful subsequent waves culminating in the so-called herd immunity goal.
All these scenarios seem to completely disregard the eventual availability of pharmacological treatment for the disease, which may improve (little or much, we'll see from the trials, no miracles though) the clinical management and thus the pressure on the hospitals.
Australia has got lucky and you should be grateful. It will largely avoid the severe and deadly over-reaction that the rest of the world has engaged in.
Firstly, given this is a coronavirus the chances of 'eliminating' it are zero unless you plan for Australia to cut itself off from the world harder than North Korea. It's infectious, it spreads. People travel. The virus will come back when your winter arrives.
But why do you even want "zeros for daily infection"? You don't chase that for the flu, and the evidence gets clearer every day that in fact, despite all the protests to the contrary, COVID-19 is no more dangerous than a bad (but regular) flu season. This wasn't apparent at first but we know a lot more than back in February. For instance it's now known not to over-use ventilators, as they can make the situation worse rather than better. In addition it's known that despite initial panic, nowhere has actually run out of ventilators, not even in Italy!
This site is an excellent collection of links with many doctors and professors stating the "similar to flu" opinion, along with links to raw data so you can check yourself:
All-cause mortality peaked at about 85,000 per week (subject to adjustment), vs the flu season of 2017 where it peaked at 70,000 a week. So COVID-19 has been, so far, about 20% worse than the flu was that year.
I guess some believe that lockdowns could push infection rates to zero. But there's no reason to believe they make much difference. Sweden's course of infection is no worse than average for Europe, and they didn't close restaurants or hairdressers or building sites. If lockdowns had the hypothesised impact Sweden should be by far leading in mortality by now but they're not, moreover most of their deaths seem concentrated in nursing homes for which lockdowns don't do much. This is also seen in Belgium (50% of all deaths in elderly care homes) and has been seen in Italy, where some nursing homes became nearly abandoned after the care workers (many foreign from eastern Europe) fled at the start of the lockdown.
Sweden is also seeing a higher impact in their migrant population for reasons currently unknown, but in the USA African-Americans are doing worse than average, so it may be skin colour related for biological reasons.
Some people are trying to distract from this by comparing Sweden to e.g. Denmark, but unlike what you suggest here lockdowns have never been proposed as a way to reduce overall death. As there's no vaccine, no cure and no way to eradicate the virus completely they can't do that. They can only push the eventual outcome forward in time, with the goal of avoiding hospital overload. Sweden's ICUs have never gone above 80% capacity, thus their strategy is de-facto successful when judged by the original justifications given in other countries.
Meanwhile, shutdowns are themselves deadly. People are told to stay away from hospitals and put in a state of fear; as a consequence some people die of other causes that could have been prevented. Additionally when the lockdowns end they face a huge backlog of operations and diagnostics that weren't done because they weren't considered as critical - some people won't get them in time and will die of treatable diseases, simply for lack of healthcare as the backlog is drained.
In the UK there's now an analysis suggesting 2000 additional people a week are dying of non-COVID related causes because of the lockdowns. There is talk of rationing of healthcare, due simply to all the delayed work and not due to COVID related load (of which there has been nearly none - overflow hospitals we...
This seems like nonsense. People in big cities rarely go less than a mile. Many are now starting to venture out into forests and similar safe areas for the first time.
On the other hand, a one-mile track in Wisconsin or Montana might be an evening drive.
32 comments
[ 32.2 ms ] story [ 119 ms ] threadIf anything I am shocked how accurate many epidemiology predictions where.
Which accurate pre-peak predictions are you referring to?
Hell the stock market still seems to think it’s no big deal.
There where plenty of different models based on different government responses, and they generally hold up well.
https://www.imperial.ac.uk/media/imperial-college/medicine/s...
https://threadreaderapp.com/thread/1239975682643357696.html
Is 1.1M is "holding up well"?
Suppression. Here the aim is to reduce the reproduction number (the average number of secondary cases each case generates), R, to below 1 and hence to reduce case numbers to low levels or (as for SARS or Ebola) eliminate human-to-human transmission. The main challenge of this approach is that NPIs (and drugs, if available) need to be maintained – at least intermittently - for as long as the virus is circulating in the human population, or until a vaccine becomes available. In the 3 case of COVID-19, it will be at least a 12-18 months before a vaccine is available . Furthermore, there is no guarantee that initial vaccines will have high efficacy.
That’s the current UK strategy and they don’t have exponential growth.
p16: "In the most effective mitigation strategy examined, which leads to a single, relatively short epidemic (case isolation, household quarantine and social distancing of the elderly), the surge limits for both general ward and ICU beds would be exceeded by at least 8-fold under the more optimistic scenario for critical care requirements that we examined. In addition, even if all patients were able to be treated, we predict there would still be in the order of 250,000 deaths in GB, and 1.1-1.2 million in the US."
We achieved better results than the best case estimate for the US in that paper, and not by a little.
But for now suppression is holding steady in the US and UK. With current rates projected out for 18 months being lower than the predicted mitigation strategy.
That said, there seems to be significant pressure to reopen, so we are likely to experience the mitigation strategy starting in a few months.
That prediction exactly matches observed outcomes. Do you expect them to predict our public policy and it’s date of implication?
Quarantine is harmful, and only useful if we’re preventing something worse. That’s probably the case, broadly — but the policies in other countries raise questions about US policies.
This is compounded by the arbitrary nature of closures, eg public construction and landscaping can continue but private can’t. People resent politicians arbitrarily (and self-servingly) picking winners and losers in the economy.
It’s further compounded by the fact that leaders haven’t announced a clear vision for when this ends. People complied because they cared and wanted to prevent deaths — but now we’re in a situation where the workforce is devastated and being hurt by arbitrary requirements, with no clear plan.
Naturally, people aren’t just going to sit there hurting — and if leaders don’t articulate a clear direction towards fixing things, people will find their own.
In other words, what have we accomplished by locking everything down? Delaying large numbers of deaths by a few months?
Extending the lockdowns until we maybe have a vaccine in 2021 does not seem like a great idea.
There are benefits to getting to 80% herd immunity sooner than later, like saving our economy and way of life.
What does that even mean ? It's slowly going down everywhere with serious lockdowns instead of going exponentially up. Anything less than exponential is a _dramatic_ decrease.
The idea that COVID-19 would grow exponentially starting from zero until a saturation of 100% was reached was only ever the outcome assumed by simulations, which have - as far as I can tell - a perfect track record of total failure. At least the sort of models created by universities in the modern era do.
To make this clear, the UW model driving much US policy response was updated several times in real time yet despite that managed to make wrong predictions for the exact same day it was updated.
https://twitter.com/alexberenson/status/1247010068148731904?...
And not just wrong, but reality fell far outside this models own uncertainty bounds too, so according to the modellers observed reality was not just unlikely but almost completely impossible.
https://twitter.com/alexberenson/status/1246465515704463360?...
(note how wide the uncertainty bounds are, yet that it's still wrong)
More commentary:
https://twitter.com/alexberenson/status/1245416486879051778?...
https://twitter.com/AlexBerenson/status/1248640413835108355
The Imperial model that drove the UK ignored the possibility of hospital capacity increases entirely, and comes from a team that has (as far as I could find out) never successfully predicted an epidemic.
Real diseases don't play out like unending exponential growth until the whole world is infected - they go into decline far before that.
Quite why this is so is unknown. The dark secret of epidemiology is nobody really understands viral spread. Every model is wrong and nobody can explain why. Germ theory must have a major gap, so what is it? My guess is a combination of much higher asymptomatic spread than currently understood (asymptomatic infection not being a phenomenon that's really been investigated much as only academic modellers care about it), and perhaps some forms of natural immunity to viruses that aren't based on antibodies.
Whatever the cause, it'd be wrong to compare the observed outcomes vs the predictions from models that are known to be completely useless and say, there's been a decrease.
https://www.nytimes.com/interactive/2018/12/10/business/loca...
We need this data now, but let’s also not forget this area needs regulation (that among other things will give teeth to violations of “but it’s all encrypted/anonymized!” claims)
https://news.ycombinator.com/item?id=20141283
https://news.ycombinator.com/item?id=22072801
https://news.ycombinator.com/item?id=21481197
https://news.ycombinator.com/item?id=21249878
https://news.ycombinator.com/item?id=19815778
The Coming Civil War over General-purpose Computing: https://www.youtube.com/watch?v=gbYXBJOFgeI
While many places in the world cannot yet justify the costs of elimination, a country like Australia, bordered by water, could have reached the goal before fatigue set in. As it is, I am resigned to awful subsequent waves culminating in the so-called herd immunity goal.
https://www.scmp.com/week-asia/opinion/article/3081110/coron...
Firstly, given this is a coronavirus the chances of 'eliminating' it are zero unless you plan for Australia to cut itself off from the world harder than North Korea. It's infectious, it spreads. People travel. The virus will come back when your winter arrives.
But why do you even want "zeros for daily infection"? You don't chase that for the flu, and the evidence gets clearer every day that in fact, despite all the protests to the contrary, COVID-19 is no more dangerous than a bad (but regular) flu season. This wasn't apparent at first but we know a lot more than back in February. For instance it's now known not to over-use ventilators, as they can make the situation worse rather than better. In addition it's known that despite initial panic, nowhere has actually run out of ventilators, not even in Italy!
This site is an excellent collection of links with many doctors and professors stating the "similar to flu" opinion, along with links to raw data so you can check yourself:
https://swprs.org/a-swiss-doctor-on-covid-19/
But you can also just look at the mortality data for Europe:
https://www.euromomo.eu/
All-cause mortality peaked at about 85,000 per week (subject to adjustment), vs the flu season of 2017 where it peaked at 70,000 a week. So COVID-19 has been, so far, about 20% worse than the flu was that year.
I guess some believe that lockdowns could push infection rates to zero. But there's no reason to believe they make much difference. Sweden's course of infection is no worse than average for Europe, and they didn't close restaurants or hairdressers or building sites. If lockdowns had the hypothesised impact Sweden should be by far leading in mortality by now but they're not, moreover most of their deaths seem concentrated in nursing homes for which lockdowns don't do much. This is also seen in Belgium (50% of all deaths in elderly care homes) and has been seen in Italy, where some nursing homes became nearly abandoned after the care workers (many foreign from eastern Europe) fled at the start of the lockdown.
Sweden is also seeing a higher impact in their migrant population for reasons currently unknown, but in the USA African-Americans are doing worse than average, so it may be skin colour related for biological reasons.
Some people are trying to distract from this by comparing Sweden to e.g. Denmark, but unlike what you suggest here lockdowns have never been proposed as a way to reduce overall death. As there's no vaccine, no cure and no way to eradicate the virus completely they can't do that. They can only push the eventual outcome forward in time, with the goal of avoiding hospital overload. Sweden's ICUs have never gone above 80% capacity, thus their strategy is de-facto successful when judged by the original justifications given in other countries.
Meanwhile, shutdowns are themselves deadly. People are told to stay away from hospitals and put in a state of fear; as a consequence some people die of other causes that could have been prevented. Additionally when the lockdowns end they face a huge backlog of operations and diagnostics that weren't done because they weren't considered as critical - some people won't get them in time and will die of treatable diseases, simply for lack of healthcare as the backlog is drained.
In the UK there's now an analysis suggesting 2000 additional people a week are dying of non-COVID related causes because of the lockdowns. There is talk of rationing of healthcare, due simply to all the delayed work and not due to COVID related load (of which there has been nearly none - overflow hospitals we...
On the other hand, a one-mile track in Wisconsin or Montana might be an evening drive.
Unfortunately walking trails were not built for social distancing.
Living in Brussels or Manhattan I rarely go more than a mile, why would I.
While in Montana, it could be a mile down my driveway to my mailbox.