9 ms·
As a physicist, it still feels odd to me they are talking about statistics with such confidence here. It seems only SK has earned that right and has tested eno
by mensetmanusman 7y ago
As a physicist, it still feels odd to me they are talking about statistics with such confidence here.
It seems only SK has earned that right and has tested enough people to get reliable statistics. If you only measure the sick in the hospital...
- soVeryTired 7y agoThe pattern is fairly consistent across countries [0]: [0] https://www.ft.com/__origami/service/image/v2/images/raw/http%3A%2F%2Fblogs.ft.com%2Fthe-world%2Ffiles%2F2020%2F03%2Ftrajectory-mar14.png?fit=scale-down&source=next&width=700 https://www.ft.com/__origami/service/image/v2/images/raw/htt...
- Animats 7y agoThat is the most useful graph on this subject so far. Go look at it now. Most countries have exactly the same growth trajectory measured from 100 cases. That's solid data about how this behaves. Singapore and Hong Kong have been able to slow things down by strong measures, but it's still a straight line on a log graph, just with 1/3 of the growth rate. Italy is on the same line, just 8 days ahead of Europe and 13 days ahead of the US. Only South Korea shows flattening so far.
- TeMPOraL 7y agoI still don't get how people - including apparently some governments - don't get or don't believe it. Getting this graph is as simple as 1) taking the daily measurements, and 2) plotting them on log scale (and technically 3) shifting the lines to overlap, but that's just cosmetics). The trajectory of the entire West being the same as Italy's is immediately apparent.
- Animats 7y agoIt's much worse than that. Italy's health system is already collapsing at 20,000 victims. The population of Italy is 60 million. That has to level off at some point due to saturation, but that's at somewhere in the millions or tens of millions.
- TeMPOraL 7y agoWorse still, Italy has one of the better healthcare systems. This doesn't bode well for the rest of Europe.
- disgruntledphd2 7y agoWell Lombardy does (number 2 in the regions of the EU). I believe that some Italian commentators have said that one of the biggest reasons for the total country lockdown was that the health systems of Southern Italy were much much worse than Lombardy, and would definitely not be able to cope.
- riffraff 7y agoFWIW, at the moment it's only lombardy's health system which is collapsing, because most of the victims are concentrated there and the growth rate there was _much_ faster because the lock down was only imposed late. The growth rate is the same in other regions but delayed, and there is some hope that the rest of the country might be able to cope with it better if the lock down measures are effective.
- sgc 7y agoHopefully. But the health care system in Lombardia has to be the best or at the top of entire the country. Southern Italy will struggle with fewer cases.
- jamesrcole 7y agoIn part, because they think they already understand the situation, and the nature of that understanding precludes taking things like that seriously. “It’s just the flu” “It only kills 2% of people” “The flu, cancer, car accidents, malnutrition, etc have killed far more people” “People were worried about SARS, Swine Flu, and H1N1, and they turned out to be hardly kill anyone.” “There number of cases here is small. Why are you panicking over nothing?” “People need to stop panicking. People are going hysterical. The media is trying to make people panic. Just look at them hoarding toilet paper”. “People are trying to ruin the economy over nothing” “The experts — epidemiologists — say there’s nothing to worry about” (I’ve had this said to me and I was given a link to a news article that spun statements from some epidemiologists to make it sound that way. They were saying that they didn’t know how the situation would pan out, and this was presented as if it meant they didn’t think there was any issue). I imagine that most people aren’t aware of the graphs. I imagine most people would be dismissive of them for reasons like the following: They simply believe the apparent normalcy in front of them more than some abstract maths. They don’t grasp the notion of exponential growth, so it seems fanciful. They also don’t understand how it’s possible to project into the future like that. Maybe they think it’s just fear-mongering speculation. What concrete evidence do they actually have for the situation in Italy? And why should they believe what’s going on in a different country will apply to their country. Maybe their country has a good health care system, and they assume that Italy doesn't.
- VladimirGolovin 7y ago> I imagine that most people aren’t aware of the graphs. And even if they are, they aren't aware of log scale.
- Frost1x 7y agoDefinitely. Noticing it's log scale (or anything non-linear) takes a trained eye. I've worked with, presented, and consumed a lot of scientific data over my career and I still have a tendency to assume plots are linear no matter how many times I tell myself to always check scale first.
- everybodyknows 7y ago>log scale There's your answer. To 90% of elected officials, a "log scale" sounds like something a timber company might weigh their product with.
- chu4n 7y agoBecause of normalcy bias: https://en.wikipedia.org/wiki/Normalcy_bias https://en.wikipedia.org/wiki/Normalcy_bias "Normalcy bias can cause people to drastically underestimate the effects of the disaster. Therefore, people think that they will be safe even though information from the radio, television, or neighbors gives them reasons to believe there is a risk." People just don't want to believe a disaster is about to happen. They will grasp at straws to explain why everything is going to be fine. That's usually not a problem because disasters are rare, but it will catch us under-prepped when an actual disaster happens.
- dfox 7y agoI have one issue with or maybe just question about this graph. Why it starts with 100 detected cases? It seems like arbitrary round number, but it also seems to be the smallest arbitrary-looking number that makes SK overlap other countries. Why is that? Why is the initial slow spread in SK irrelevant? What caused it to be initially slower than in other countries?
- contravariant 7y agoThis doesn't really matter, you're just doing that to make it easier to compare. Shifting South Korea so it doesn't overlap the other countries is possible but not terribly useful. They do seem to be missing Denmark though, which should stick out somewhere top-left of Norway.
- Animats 7y agoThey do seem to be missing Denmark though, which should stick out somewhere top-left of Norway. I was just talking, via text, to a friend in Norway. They're going for major restrictions on contact and the people there are not worried enough yet. I sent her that graph.
- loopz 7y agoNorway is going into full lockdown (stay at home, don't hoard). Borders closing except for export/import from Monday. Hut vacations now getting banned as well.
- KozmoNau7 7y agoWhich is a good plan and similar to what Denmark is doing, ie. stay at home as much as possible, everyone who can work from home does so, closing the borders for all non-essential traffic, and so on. There is a whole system being set up for compensation to concert venues and other businesses who are severely hit by the lockdown and blanket ban on gatherings of more than 100 people. Open air walks are ok, as long as you keep a reasonable distance to other people. Only go shopping if it's absolutely necessary. And yet some people are still going to crowded bars, like nothing's going on. It's amazing how stubborn some people are.
- mensetmanusman 7y agoThis could also be a measure of the rate of test kit production by country.
- Fnoord 7y agoThe graph can be reproduced with more recent data here [1]. It is based on WHO data, and lags a bit behind. If you want to see more data and graphs, see [2]. [1] https://ourworldindata.org/grapher/covid-confirmed-cases-since-100th-case https://ourworldindata.org/grapher/covid-confirmed-cases-sin... [2] https://ourworldindata.org/coronavirus https://ourworldindata.org/coronavirus
- davidw 7y agoNo one is 'confident' about a lot of things. But everyone who has seen what is happening in Italy is confident that taking a 'wait and see' attitude will cost lives.
- bsaul 7y agoi wish more countries did randomized testing over its whole population, or extensive testing, like SK.
- wnkrshm 7y agoExactly. I don't see how trying to limit testing to people who had contact with known cases or traveled to at risk countries (as practiced in Germany) can be successful.
- FiberBundle 7y agoIf the availability of testing kits is low and the ratio of infected to non infected people is extremely low, it makes absolutely no sense to sample randomly from the population.
- lucb1e 7y agoHow many do you need though? If you take 0.1-1% of your testing capacity, those handfuls of samples a day give you some information. Not regional level or very reliable data, but it gives an idea of whether people (as a population) are having it and are having mild symptoms and are not reporting anything, or if it isn't widespread yet and the hospital cases are the only cases.
- gpderetta 7y agoAldo if you expect the true positves to be a small percentage (which should still be the case in most countries), false positives will dominate in a ranom sampling. Apparently false positives are high for unsymptomatic cases. Of course if you can estimate the false positive rate well enough, the random sampling might still give useful statistics.
- bsaul 7y agoAs someone who's dealing with analytics data regularely, thank you ! i absolutely don't understand how can someone give any ratio about this disease knowing how only the most severe cases got tested ( except in SK), and that the symptoms are almost undistinguisable from the flu.
- wnkrshm 7y agoWhat do you think of the practice of only testing people who need hospital-level care and people who had contact with confirmed covid-19 patients (and people who traveled to at risk regions)? I would think that this can only give you the infected from a few clusters and you can be surprised by an unknown cluster of patients suddenly needing intensive care. Edit: And due to the long incubation period, that cluster can be quite large.
- marcosdumay 7y agoTesting the entire population is not viable. If you don't have many in-country cases, you can delay the epidemics arrival by controlling just the people that come from elsewhere. Of course, after the disease starts propagating internally, testing becomes irrelevant, and it's time to reduce its spreading speed.
- empath75 7y agoIt simply must be much more dangerous than the flu for the simple reason that comparatively few people have it and it is overwhelming hospitals.
- bsaul 7y agoi wonder if it's really true that "few people" have it. We do observe a huge amount of people going to the hospital in italy, but that could very well be because we reached the peak of the epidemie, where almost all of the population have been infected and remain asymptomatic. in this hypothesis, you implement a lockdown and see number drops over two week but what you're really seing is a population reaching herd immunity.
- ISL 7y agoIf they're making this statement based only on the cases they know about, then the argument is even stronger if there are more cases.
- kspacewalk2 7y agoThere are more cases, they are on average much more mild, the real mortality rate is much lower, it is actually not doubling every two days (it is the tests that are doubling every few days). Lots of possible and very consequential conclusions.
- TeMPOraL 7y agoDeaths follow a predictable exponent too.
- kspacewalk2 7y agoBut with several weeks worth of lag. The number of deaths today should be divided by the number of cases N>10 days ago. The number of cases was much more severely underestimated N days ago than today. The death rate in Italy is basically unknown because the denominator is just noise with a time lag.
- loeg 7y agohttps://medium.com/@tomaspueyo/coronavirus-act-today-or-people-will-die-f4d3d9cd99ca https://medium.com/@tomaspueyo/coronavirus-act-today-or-peop... Start reading at "2. What Will Happen When These Coronavirus Cases Materialize?"
- Too 7y agoThank you. Recommend everybody to read this.
- apexalpha 7y agoWho is this mr Pueyo? And why should we listen to him regarding this subject? From his Medium profile page it seems he is a 'Storyteller' and 'Viral marketer'. So, someone who knows how to tell a story with confidence even if you know nothing of the subject. I'm gonna listen to my government regarding this, rather than medium posts by viral marketeers. Thank you.
- loeg 7y agoYour government is saying the exact. Same. Thing: https://www.government.nl/latest/news/2020/03/12/new-measures-to-stop-spread-of-coronavirus-in-the-netherlands https://www.government.nl/latest/news/2020/03/12/new-measure... https://www.rivm.nl/en/news/current-information-about-novel-coronavirus-covid-19 https://www.rivm.nl/en/news/current-information-about-novel-... > 15-03-2020 , 176 new patients, a total of 1135 > 14-03-2020 , 155 new patients, in total 959 > 13-03-2020 , 190 new patients, in total 804 > 12-03-2020 , 111 new patients, in total 614 > 11-03-2020 , 121 new patients, in total 503 > 10-03-2020 , 61 new patients, a total of 382 > 09-03-2020 , 56 new patients, a total of 321 > 08-03-2020 , 77 new patients, a total of 265 > 07-03-2020 , 60 new patients, a total of 188 > 06-03-2020 , 46 new patients, a total of 128 > 05-03-2020 , 44 new patients, a total of 82 > 04-03-2020 , 15 new patients, a total of 38 > 03-03-2020 , 6 new patients, a total of 24 > 02-03-2020 , 11 new patients, a total of 18 > 01-03-2020 , 1 new patient, a total of 7 ... Here's a log-scale graph: https://www.chartgo.com/embed.do?id=7e5c9f30aa https://www.chartgo.com/embed.do?id=7e5c9f30aa (And just for comparison, not log scale: https://www.chartgo.com/embed.do?id=7e5c9f30aa https://www.chartgo.com/embed.do?id=7e5c9f30aa ) Looks like NL has had steady exponential growth over the past week or even 10 days. As far as Mr Pueyo, in this context he appears to be an effective science writer, which is what matters. If you want a more succinct source from an academic, there's http://ctbergstrom.com/covid19.html http://ctbergstrom.com/covid19.html .
- trashtester 7y agoAs a physicist, you are familiar with formulas similar to this: N_obs = p_detect * 2^( n_days / T_double ) N_obs : The number of observed cases p_detect : The fraction of actual cases that are detected (assumed to be constant) n_days : number of days T_double : The number of days it takes for the number of real cases to double If you examine the formula, the p_detect constant is almost insignificant. If T_double = 2, a difference in a detection rate of 0.2 and 0.8 is only 4 days for n_days. As long as T_double is less than several months (and constant), an absolute catastrophy will occur. Within a weeks from now, nearly the entire fraction of the population that is receptive to infection will get infected, and will be sick basically at the same time. With 10-15% of the population requiring intubation, we can end up with a significant part of the population in each country ending up dead.
- ufov2 7y agoThat assumes you are sampling uniformly, and not in already saturated clusters. The doubling time is also not a known constant, as it depends on actions taken etc, and uncertainty in it will broaden the distribution. Obviously, all this is basic statistics and should be known to epiodemiologists, who hopefully have some input to policies. For large N_obs, this will maybe be less important as you're going to find the severe cases anyway, but the uncertainty is significant in the beginning stages, and it is unfortunately these stages where actions will have most impact.
- trashtester 7y agoAs long as growth remains linear in an exponential plot, you are nowhere near the saturation point. When about half the population is already infected, the number of new cases should grow roughly linearly. In Italy, it seems to still grow exponentially. Itally has already passed the number of simultanous cases that their healthcare system can handle. I don't believe they really want to find out at what level the saturation point can be found. That could mean hundreds of thousands, potentially millions, dead.
- alexandercrohde 7y agoThe point is that the mortality rate estimates are going to be totally out of whack without accurate measurements. As well as estimates of the currently infected. This matters a lot. The first number is obviously important. The second matters so Italy can plan for whether things have peaked yet, or how many weeks out until it does.
- jiggawatts 7y agoSo, as a physicist, you should know how exponential growth works. Let's say you have a pile of Uranium that's juuuust sub-critical. You would likely to be confident to stand next to it without significant fear. But if someone throws on just enough additional Uranium for the pile to go supercritical, you would run for your life because none of the following will matter[1] all that much to the final outcome: 1) How much neutron radiation there was to start with. 2) The specific exponential growth coefficient. 3) How much total Uranium there is, as long as it's a nontrivial amount 4) Whether your Geiger counter is off by some constant factor or not. None of those matter. It's going boom. It might go boom a bit sooner or a bit later, but it is going to go boom. That's just how exponential growth works. There's no maybe. There's no "let us wait and see". There's no "we'll hope for the best". No: It. Will. Go. Boom. All of the factors that are irrelevant are just shifting a figurative vertical wall on your graph paper a bit to the left, or a bit to the right. It's adding or subtracting "a couple of doubling times". If the doubling time is short, then you're not really achieving anything by fiddling with constant factors. With the Coronavirus just about every country has a doubling time of 4-5 days. I live in Australia, where we're about a "month behind" everyone else. So of course, the dumbass government is saying things like "the heat here will slow it down". But it hasn't, our doubling time is 4.75 days at the moment. They're saying that it's "premature to lock down the country". No it isn't, it's already weeks too late and getting literally exponentially worse daily! They're saying that the hospitals are being prepared, but no amount of extra beds or ventilators will help. If they double the number, it just delays the catastrophe by 4.75 days. Not even a whole week! Quadrupling beds and ventilators buys just under 10 days before people are being turned away to go home and die. So, again. I ask you: If you were the nuclear physicist in charge of a nuclear pile and someone told you it's gone critical and the radiation is rising exponentially, is your first reaction to: just "stand around and wait and see what happens", or to: take drastic action right this second? Do you SCRAM, or do you call the communist party leadership for permission? Are you the hero of the Chernobyl story, or the villain? [1] Ignoring thermal effects slowing down the reaction. This actually has an analogy with disease spread, where there are fewer susceptible people remaining because everyone is already sick or dead. If we reach this point, we'll be seeing millions infected and hundreds of thousands dead.
- ghthor 7y ago
- eyegor 7y agoIf you make a bunch of assumptions about the distribution, you don't need a large sample size to make decent estimations about the population. Of course, the only way that works is with random or stratified random sampling. With reporting bias, I'm not sure any country has enough data to make statistically confident statements with any precision. Folks with a stronger statistics background are welcome to correct me.
- dguest 7y agoIt's interesting that you lead by saying "as a physicist". As a physicist I'm used to thinking about orders of magnitude and whether something grows like log(N) or N!. I'm pretty sure all the data points to this being O(a big problem) but I'd leave it to the epidemiologists to say anything with confidence.