Specifically, (which I mentioned in the comments) how can the peak be less bad and quicker than expected? If social distancing works, then the peak should be delayed. If social distancing isn't working, then it should be tearing through the country.
I am going on a bit of half-remembered and hastily revised undergraduate epidemiology here, but I don't think that's right. You could definitely have a situation where an intervention brings the peak forward and decreases the peak value.
Take the ad absurdum situation: we have a pathogen that incubates for one day and then makes people obviously sick. We are running at 50 cases a day today, and we look at the epidemic model and predict a peak of 1000 cases a day in three weeks' time. Let's say that we have complete information on the pathogen and the structure of the population, so this is a completely accurate prediction, absent any intervention.
Immediately we impose total social distancing: everyone goes into a little room by themselves and doesn't have any potentially infective contacts with anyone else. Within one day, the number of new cases has dropped to zero. We've brought the peak forward - the peak was expected at T plus 21 days, and thanks to the intervention the peak happened at T plus 1. And we've made it lower - we were expecting 1000 cases a day at peak and we got 50.
how can the peak be less bad and quicker than expected?
I think the problem here is the assumption of a symmetric curve with a fixed area under the curve. If you assume the curve is necessarily symmetric, "flattening the curve" is going to push the peak off into the future and reduce the slope on both sides. But that's just an assumption (in fact, more of an aesthetic design choice).
What people are actually doing by social distancing, though, is preventing the virus from spreading - drawing a completely new curve, if you like. If each case gives it to one other person, the cumulative cases will grow linearly and the total at any one time stay the same as equal numbers of people get it and either die or recover. If each one gives it to more than one, they accelerate and it will grow. Once it's less than one, it will slow down and eventually begin to shrink.
If the distancing works well, it will *stop* the process that the curve is measuring and obviously when you stop a process that's growing, that's the peak.
(The specific IHME model has done some really weird stuff like estimating the UK was already over ICU capacity by 700-odd at a moment when there were that many beds available in London alone.)
I suppose. So the models end before social distancing is ended and life resumes? They only take outbreak at a time?
Anyway yes, social distancing works. If you half-ass it your curve pops right back up again. (Incredibly long and shittily-written piece.)
7 also yet another score against benign dictatorship.
For values of "we" that include all of Americans, we're going to half-ass it. There's not going to be enough support for people to stay out of working for long enough even among the people who want to.
You were supposed to make a cock joke there.
2 is a good way of putting it except for this bit: "If you assume the curve is necessarily symmetric, "flattening the curve" is going to push the peak off into the future and reduce the slope on both sides. But that's just an assumption (in fact, more of an aesthetic design choice)."
It isn't an aesthetic choice - it's a product of the model. A simple SIR model predicts infections per day based on beta xy - beta is the number of potentially infectious contacts per day, x is the fraction of infected (and hence infectious), y is the fraction susceptible. (R is beta y times the length of time t that you stay in the infected phase; Ro is beta t because at the start of the outbreak y is one.)
If you run that model, it will show a symmetrical bell curve for infections per day, and a classic s-shaped logistic curve for total cases over time. It'll be slow at first because you've only got a few people going round infecting people; then it'll accelerate as more people get infected and pass it on, increasing x; then it'll level off again, because all the time y is declining as susceptibles turn into infecteds and then into recovereds.
But the key point is that this only works if beta stays the same throughout. If you change beta then you are, as Alex says, drawing a different curve. If you drop beta enough that R goes below one, you'll peak instantly and the epidemic will end with a lot of the population still susceptible. If you drop beta a bit less than that, then the epidemic still burns through most of the population (one minus one over Ro) but the peak will be further off and a smaller fraction of the population will get it - the area under the curve will be less.
A friend of mine lost his father to Covid-19 (probable) yesterday. The man was in a care home and the disease just tore through the whole facility. I don't know if they didn't try or weren't allowed to transfer anybody to a hospital, but care was only done in the facility where he resided. And I'm wondering if there's any reason to assume Louisiana care home are going to be worse than other states?
All those "flatten the curve" diagrams everyone was passing on were wrong because they showed a symmetric curve. The only thing that could "flatten the curve" like that is if the first curve was wrong - based on faulty parameters. Even then the area under the curve would be different for each one because the area under the curve of cases per day is total number infected, which is total population times (one minus one over beta t) and so if you've got your beta or t parameters wrong then that's going to give you a different area, isn't it?
If you drop beta enough that R goes below one, you'll peak instantly and the epidemic will end with a lot of the population still susceptible.
Right, this is our goal. Followed by Phase B: use contact tracing, widespread testing, isolation of exposed people to prevent another peak. Or in the US, just send people back to work to get sick because they can't afford to stay home anymore, with contact tracing sort of a thing you can sign up for if you want, and isolation being handled by staying home to infect everyone else in your home.
My local area is now mandating people wear masks in stores. Sounds good if they provide masks to people, which I can't even imagine happening. They are also mandating stores provide masks for employees if they can get Jared Kushner to sell them any. I hope the free market, in the form of people on Etsy shifting to making fabric masks all day, can keep up.
I somehow thought that when it was declared that containment was no longer possible, it excluded the scenario in 2. That we were modeling until herd immunity or vaccine.
My local area is now mandating people wear masks in stores. Sounds good if they provide masks to people, which I can't even imagine happening.
Or people could just make their own. Needles and thread are cheap. Scrap fabric is free. Everyone can sew or has a family member who can sew.
The Czechs made like ten million masks for themselves in a weekend.
The Czechs have like 10 million COMECON-trained grannies.
15: "herd immunity" is a moving target, though. It isn't a set value for one pathogen, it's a set value for one pathogen and one particular structure of population and how often they contact each other.
To get it, you need (one minus one over Ro) of your population to be immune. But - see above - Ro depends on beta and t, and beta is exactly what you're altering by putting lockdown etc in place. You could easily have a situation where you get "herd immunity", as in the epidemic burns out, with 80% of the population still susceptible, if t was five days and beta was 0.25. (Check the maths.)
But then you lift lockdown! And beta goes from 0.25 to 4! And suddenly you don't have herd immunity any more!
All it means to say "we have herd immunity" is "if you put an infected person in our population, they will on average infect less than one other person". The population in the absurdum scenario in 1 has herd immunity even though none of the people are immune, because no one contacts anyone because they're all in separate little rooms!
19 is pedantic in the extreme or wrong.
I prefer to think of them as being in yurts. But that doesn't alter the underlying maths.
21: it's an absurdum example designed to highlight how the percentage of recovereds in the population required for herd immunity can vary dramatically depending on beta.
But if 'immunity' means anything, it would have something to do with transmission being blocked because a potential host was immune.
Nothing means anything! Everything is permitted!
17 is excellent. and true! we are relying on the etsy connection, first deliveries tomorrow, and in the meantime the folded-hanky no-sew technique. although i never could get any hair bands to work, so we are using strategically-knotted trimmings from worn out fishnet stockings, they work fine except for it is tricky to prevent the folded hanky models from steaming up your glasses.
newsom, who i still regard as one of the more slippery characters i have ever personally dealt with! very much do not trust the guy!, managed to reassure me with his press conference yesterday. if he can harness his galactic sized ambition and major slipperiness to evade trump actively and viciously undermining the west coast states (a huuuuuuge if), then we have some chance of limping through this out here.
at some point we are going to have to grapple with the year-over-year increase in the death toll, as i'm convinced there is now and is going to be until our bright shining universally administered vaccine future an enormous number of deaths not officially counted as but overwhelmingly likely due to the plague, per 12. am very apprehensive about my 94 yr old mother in law, if it gets into the building she lives in (apartments, not care home, but all old folks), it is going to scythe viciously.
"Herd immunity" is just another way of saying "beta yt in this population for this pathogen is less than one". Making beta equal zero is one way of achieving that. Making beta small enough (social distancing or whatever) is another. Making y small enough (ie vaccinating lots of people, or having had lots of people catch the disease in the past) is another. And reducing t is a fourth- if you have a bacterial disease, say, treating it with antibiotics cures infecteds faster, so they get to pass it on to fewer people.
If symmetry is a baseless expectation for a curve, then what's a more realistic shape? That it trickles down more slowly than the increase, with sine waves for recurring outbreaks? That it plummets because we're in isolated little rooms?
Isn't Ro by definition the transmission with no preventative measures, and R the transmission with distancing, etc.?
That's needlessly confusing. There are two ways to reduce transmission, blocking contact and development of immune responses. You need different words for both even if you bake a simpler model with only one term.
30 to 27. In other words, if you define 'herd immunity' as you are doing, you have not only changed the usual definition of the term (which has always been used to refer to people developing an immune response), but you've covered over a crucial distinction for policy making/behavior.
A simple SIR model predicts infections per day based on beta xy - beta is the number of potentially infectious contacts per day, x is the fraction of infected (and hence infectious), y is the fraction susceptible.
I agree with Moby. Herd immunity should deal with Y, social distancing affects beta.
29: not quite. Ro is the initial rate of transmission - as in, one infected person in the population, and everyone else susceptible. As the epidemic goes on, R will drop, because infected people will have a growing chance that some of their contacts will be with other infecteds or recovereds, and therefore won't produce new cases; but at the start of the population, everyone they contact catches the disease. Ro with social distancing will be lower than Ro without it.
28: really no way to say; it depends very much on how much beta has changed as a result of precautions, and how big the proportion of recovereds is. You get a sort of sawtooth wave (not really a sine wave) if you do an on-off strategy; you relax precautions, and then when case rates go above a certain level you put precautions back on again, and so on. And the spacing of the peaks and the slope changes as you go on, because your proportion of recovereds is rising each time.
You get a sine wave if you include birth and death in your model, because everyone is born susceptible. And there are a lot of diseases that show this - you get a measles outbreak every few years because that's how long it takes to build up a new population of susceptibles that's big enough to push beta y t up above 1 again (by pushing y up).
Herd immunity should deal with Y, social distancing affects beta.
But the value for y that you need to achieve herd immunity depends on the value of beta. Immunisation deals with Y. Herd immunity is the result, not the input.
It's remarkable how quickly our impressions of what's going on, and what the appropriate mental models to use changes.
I have the feeling that I spent a while mentally preparing myself for, "this is going to keep getting worse, quickly, for the short-term, and is going to be a major problem for the next 12-18 months until a vaccine is available." Then, to switch, within the last 2-3 weeks, to thinking, "I'm surprised, the interventions appear to be working better than expected. This is bad, but I think we've clearly avoided the worst-case scenarios." But both steps have actually happened fairly quickly (on the order of weeks), and it's a reminder of how much uncertainty there is, and how little reliable data we have.
The fivethirtyeight article illustrated with cartoons about the challenges of modelling is very good.
To the extent I have a take-away a the moment it is -- social isolation appears to be effective; even more effective than I would have guessed when the orders were first being made. Therefore we should continue doing what we're doing until we have clear signals that we can relax. Optimism is most productively expressed as, "be attentive to discipline and the rules of isolation and there's a chance that we can get to a point within the next month or two where we have multiple days with 0 new recorded cases." (almost certainly true in my county; maybe true in my State, not true in the US as a whole).
I don't know if that's the best (or accurate) take-away, but there's my read of the numbers.
If you have no transmissions because of social distancing, you do not have herd immunity. You need a sufficient proportion of the herd to be immune before you have herd immunity. That's why they use the words 'herd' and 'immunity' next to each other like that.
You need a sufficient proportion of the herd to be immune before you have herd immunity.
I'm with Moby Hick here. If you have a disease that is sufficiently difficult to spread that, even in a population without immunity it is unlikely to expand beyond a small circle (very low beta) I wouldn't call that heard immunity, I would call it a not-very-infectious disease.
"The most similar virus to this shows there is not a protracted immune response," Walt said.
36: I mean, OK, if you want. I did say it was an absurd example. I'll say it again if you like. It's an absurd example. I'm taking the situation to a ridiculous extreme to highlight what I think is an important feature of the way epidemics spread and the way in which herd immunity works. Nothing like this would ever happen in the real world. I'm sorry for even bringing it up.
26.3. I wonder if the increase in death toll will be partially counteracted by the decrease in traffic accident deaths. Will there be more or less murders? So much interesting research fodder for social scientists.
All the data at this point is so suspect -- the tests are wrong about a 1/3 of the time - a death is a death, but how many are Covid-19 related? A lot of the time this determination is fairly arbitrary, I would guess. In these circumstances, if the data fit the model well, that would be a good reason to suspect someone is cooking the books.
It is an important feature. But herd immunity is one possible outcome (depending on how long immune responses last and vaccine development success and the like) and not a synonym for the end of the epidemic (which could also happen from isolation). Those are very different outcomes.
Good thing a Leninist government didn't cook all the initial stats, burn the bodies and destroy the samples.
I also have questions about the tests being wrong 1/3 of the time. My understanding is that the test itself is quite sensitive, but it's got significant user error - not swabbing deep enough, or the person isn't shedding virus in that particular spot, or whatever.
Take my county: There have been 750 tests, of which 100 were positive and 630 were negative. So of those 630, if a third are wrong, that's another 210 cases.
I am very sure that no one is being tested here without significant symptoms - most people I know who have symptoms aren't getting tested, and there's no drive up testing or anything convenient like that.
So what do the other 420 people have, who are still negative? They probably don't have the flu, because it's already April and social distancing must be reducing the flu significantly.
It makes me suspect that the 1/3 false negative is too low, and that it's even higher than that.
I wonder if the increase in death toll will be partially counteracted by the decrease in traffic accident deaths.
Well, "In the average week in San Francisco, California, since the start of 2019, there are typically about 260 calls made to 911 about traffic collisions. For the week of March 17-23, there were just 110" - https://qz.com/1822492/traffic-accidents-are-plummeting-because-of-the-pandemic/.
And "Coronavirus stay-at-home orders that went into effect on March 20 have reduced vehicle collisions on California roadways by roughly half, according to a UC Davis survey that is the first to estimate the impact of the extraordinary health orders on traffic." https://news.yahoo.com/coronavirus-stay-home-orders-reduced-130040516.html
California loses 3500 people a year to traffic accidents. So looks like keeping California locked down for two months would save about 300 lives. The epidemic's forecast to cost 1483 lives.
But then we need to take into account how many additional people will die due to boredom-induced novel sex acts.
I'm pretty sure kids are taking the lower traffic levels as an opportunity to race motorcycles down city streets. Our maybe it's just quiet enough to hear them now. I don't think that will result in as many traffic deaths as usual, but I'm guessing it will result in some really messy ones.
my county: There have been 750 tests, of which 100 were positive and 630 were negative.
And you're still wondering if flyover is going to get schwacked?
46: I mean, it's California, so I assume that around 60% of road traffic accidents are caused by boredom-induced novel sex acts even under non-pandemic conditions, and the rest of them are caused by high-speed police car chases.
44: Is that rate of negatives possibly because of worried-well people bring more about to push to get a test than the people at higher risk bit maybe with less ability to get medical care? Or because they are wisely testing medical workers to see that iatrogenic cases stay lower?
48 Yeah, my Trump supporting dad was still denying that most of the US is going to hit hard when I spoke with him a few days ago, i think most places that aren't taking action like California or Washington are going to look a lot like New York in a couple of weeks.
100 cases and one fatality in a county of 250K doesn't feel like getting thwacked, no. Or even 250 cases and one fatality.
I'm taking the situation to a ridiculous extreme to highlight what I think is an important feature of the way epidemics spread and the way in which herd immunity works. Nothing like this would ever happen in the real world. I'm sorry for even bringing it up.
I'd put it differently. We're all trying to gain an amateur familiarity with what mental models are required to think about epidemics. I think of these conversations as attempting to build mental vocabulary ("I've been using this model, or this language, or these categories in my head to help make sense of things."). We're not in a situation where we have a shared understand and we're trying to figure out how to apply it (or argue about the finer points of it).
So, I'm curious, my initial reaction to your comment was, "I'm not sure how this helps me build my own mental model." So rather than giving up, is there another way to explain what the feature that you're trying to highlight is, and why it's important?
i think most places that aren't taking action like California or Washington are going to look a lot like New York in a couple of weeks.
One of my reasons for optimism is that, two weeks ago, I thought LA and FL were both going to get very bad, very quickly and it doesn't appear to have happened (yet). I'm not quite sure why, but it isn't looking as bad as I feared.
You know what's definitely down? Elephant maulings.
Last November, a wild elephant killed a 52-year-old man after pulling him out of a bus near Corbett Tiger Reserve in Mandal forest range in Nainital district.
So rather than giving up, is there another way to explain what the feature that you're trying to highlight is, and why it's important?
Sure - basically, the key to the end of an epidemic is to have each infected person give rise to an average of less than one new infected. Then the epidemic will gradually die out - each 'generation' of infecteds will be smaller than the previous one. To produce a new infection, an infected person has to
a) contact someone else
b) who is susceptible to the disease
c) while they themselves are still infectious.
Quarantine attacks point c): if you're infectious, we shut you away where you can't contact anyone else, until you're not infectious any more.
Social distancing and other precautions - like condoms for STIs - attack point a) - we reduce the number of contacts people have. ("Contact" is a potentially infectious interaction. It could mean "being sneezed on" or "shaking hands with" or "having unprotected sex with" depending on the disease we're talking about.)
Vaccination attacks point b) by making it more likely that someone who contacts an infected person is actually immune, rather than susceptible. If you get the proportion of vaccinated people high enough, then the epidemic will just die out; and, if another infected person arrives in the community, they won't give rise to a new epidemic.
But the point I wanted to make is that "high enough" isn't just dependent on the disease; it's also dependent on how many potentially infectious contacts people tend to have with each other. If you are infectious for four days, and you contact one person a day, then if 80% of the population is immune, you'll only infect (on average) less than one other person. That's "herd immunity". Even though not everyone is immune, enough are immune to prevent an epidemic.
But if you are infectious for four days, and 80% of the population is immune, but you contact ten people a day, you'll infect (on average) eight other people. Same disease; same proportion of the population immune; but the population no longer has herd immunity, because its structure, the number of contacts people tend to have per day, has changed.
This is also why it's wrong to think that if you don't do anything, "everyone will get it". There's no sort of epidemic that could spread through 100% of the population.
There's no sort of epidemic that could spread through 100% of the population.
Yet.
I suppose one where infected people are immortal and remain infectious forever. But even then Will Smith manages to avoid it.
54: It's gotta be about weather, right?
I think Ajay and Moby are both sort of right in 34 vs 36. You can use "herd immunity" to mean "enough immunity to cause the disease to start dying out, given whatever beta we have managed to extraordinarily socially engineer" (as Ajay does) or "enough immunity to cause the disease to start dying out, given the beta the disease would display under ordinary social conditions" (as Moby does). The former could be considerably lower than the latter, even zero if we all stay completely isolated.
Given that R-nought (average number of people an infected would infect in a completely susceptible population) is beta.t, and some but not all authorities (per Wikipedia article on R-nought) specify "absence of 'any deliberate intervention in disease transmission'" in their definition of R-nought, those authorities are implicitly using Moby's definition of R-nought. The others are presumably using Ajay's. I think Moby's fits ordinary linguistic intuition better.
s/b "given whatever beta we have for the disease, even if we have had to extraordinarily socially engineer it."
s/b "Moby's definition of 'herd immunity'"
I want someone to explain all this using cartoons of a herd of buffalo and Trump sneezing.
62.2: it does make more intuitive sense but it's less rigorous - including things like "absence of any deliberate intervention" leaves the way open for a lot of interpretation. Ro for flu in a cold climate society where people customarily greet each other with handshakes and embraces and travel around on crowded trams will be a lot bigger than Ro for a tropical society where they bow to each other and cycle everywhere.
I'm not denying there are lots in factors in Ro. I'm just pointing out that Ro is in no way a synonym for "herd immunity."
I think Ajay's wrong about how he's using "herd immunity" and entirely missing the point of herd immunity. The point of herd immunity isn't to make the virus go completely extinct. That would require a reproduction rate below 1 everywhere in the entire world. Experience has shown that this is extremely difficult and requires decades long campaigns. Herd immunity is about safety from new infections coming in and spreading. We're not social distancing *forever*, and so we don't have herd immunity until we hit 1-1/R0 under normal circumstances. Measles was eliminated from the US, but we still need 90%+ immunization rates to maintain herd immunity.
If you want to make this into policy you definitely need some buffalo.
We passed 400 confirmed cases last night -- out of a population of a million. Of those, more than 200 are listed as recovered. The number of new cases has been under 10 a day for several days, having been above 20 a day more than a week ago. Still at 7 deaths -- only one in the last 10 days. 21 hospitalizations. 9500 tests altogether, 350 in the last day. 5 new positives on 350 tests doesn't sound awful.
The distancing is definitely working.
Have we peaked in the curve we're on? I guess it does kind of feel like it.
65. Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo buffalo herd immunity buffaloes...
66. Fair enough.
If you drop beta enough that R goes below one, you'll peak instantly and the epidemic will end with a lot of the population still susceptible. If you drop beta a bit less than that, then the epidemic still burns through most of the population
IANAE but, super-pedantically, wouldn't you have to have dropped beta quite a bit less than that to have the epidemic burning through more than 0.5 of the population? I.e. you couldn't have reduced R-nought to less than 2 if that were still to happen. If you drop beta enough that R-nought will be just above one, e.g. 100/99, then won't the still-susceptible (hence never-infected) proportion of the population in the long run be 99%?
70 is kind of amazingly proportional to my county.
Apologies if I appear to ignore any responses - (as is often the case with me) I have to run.
I can't speak to the model debates, but isn't the basic takeaway of where we are as follows: social distancing worked better than expected, but we were supposed to use that time to ramp up testing and contact tracing, we didn't nearly enough, so now any plateau or fall in cases we see represents at best a stalemate because it's all going to start up again sooner or later?
75 is how I understand it. I mean, we haven't really been isolated enough for it to burn out around the world, if that's even possible.
Here's a thing I have been mulling over: could the US, or similar economies, reopen with people in clumps? THat is, if we could work out groups that were self-contained for social and work interactions, and have as much non-interacting economic activity as possible, would that contain COVID19? And is there any way to make such clumps that we could bear?
I think the answer to the latter is no just because there are too many couples and no good rule for `a couple's group only includes the parents/stepkids/paramour of the $UNIQUE person in the couple'. I want to see my parents, my sweetie wants to see his, but we both have sibs and now their sweeties' parents are in our group and for all I know the country doesn't actually partition at all that way. Just one big susceptible family.
My guess is that we reopen from social distancing through an uncoordinated spate of desperate people with no other means of supporting themselves and huge assholes gathering out of spite to stroke guns in public.
76.last. See this modeling for example: https://www.washington.edu/news/2020/04/13/just-one-friend-covid-19/
I don't find that totally convincing - if I were going to visit "just one friend", I wouldn't choose a household where someone was still working out of the house.
Also, is visiting "just one friend" that different from going to the grocery store?
I am not visiting any friends, but I'm also not quibbling about the margins with people who are doing the large things with vast gains.
If you visit somebody, you are visiting everybody they have ever had sex with.
Or everyone they have had sex with since December 2019.
79 -- I had questions about what counts as contact sufficient to spread infection. If I see a friend and visit with them for an hour in a park standing 8 ft away is that okay? (I have done essentially that, but only once or twice)
Thanks, NickS.
So we couldn't choose our just-one-friends, because they link up*, but we could be assigned to a, er, granfalloon of people we mostly were exposed to anyway. I expect the chance of each granfalloon having a star-crossed lover or other reckless idiot in it connects everything up anyway. OTOH, I didn't think we could do social distancing this much, maybe we could granfalloon if each 'oon were assigned a natty and very visible hat.
* It's so hard to believe that atomized racist classridden society as it looks in PRIZM is connected. Guess so!
We were independently talking about the 'just-one' thing tonight. I think it would make the isolation much more bearable, but of course it would be impossible to police and there's always one idiot who ruins it, and with this virus it seems like all it takes is one idiot. But. The families of the Calabat's two best friends have decided to do this -- the mom of one is divorced, has to run her four classes from her house, and the dad who shares custody I believe is at work on site still -- so the kids get together and play as they swap childcare. They are sensible people whom I trust to keep to their group of six. I think they're being reasonable. But. The Calabat had a virtual playdate with them today and he was so envious he was nearly physically ill. I thought we couldn't have friends because of the virus, Mom. They're playing. Why can't I play? We've been staying home. We're not sick. Worst Zoom date ever.
I've been thinking that given that COVID 19 is fairly safe for the under-40s, that it is probably worth doing variolation. Have the low risk (low blood pressure, no asthma, etc) people go to COVID camp for a month, with alert health facilities on call for the rare very sick people. But I don't see how it could scale, so maybe it isn't worth it. Could maybe use campuses/dorms for it.
The young people will just take all the good food before I can get any.
Hey, Megan, 38 is especially for you. Stick to your day job.
Did young people get all the good butter at your store too?
Or at least people who wake up before nine.
Yes, but it's fine. I just had the young people dance in circles until they melted.
85
A propos of distempers, I am going to tell you a thing that will make you wish yourself here. The small-pox, so fatal, and so general amongst us, is here entirely harmless, by the invention of ingrafting, which is the term they give it. There is a set of old women, who make it their business to perform the operation, every autumn, in the month of September, when the great heat is abated. People send to one another to know if any of their family has a mind to have the small-pox: they make parties for this purpose, and when they are met (commonly fifteen or sixteen together) the old woman comes with a nutshell full of the matter of the best sort of small-pox, and asks what vein you please to have opened. She immediately rips open that you offer to her, with a large needle, (which gives you no more pain than a common scratch) and puts into the vein as much matter as can ly upon the head of her needle, and after that, binds up the little wound with a hollow bit of shell; and in this manner opens four or five veins. The Grecians have commonly the superstition of opening one in the middle of the forehead, one in each arm, and one on the breast, to mark the sign of the cross; but this has a very ill effect, all these wounds leaving little scars, and is not done by those that are not superstitious, who chuse to have them in the legs, or that part of the arm that is concealed. The children or young patients play together all the rest of the day, and are in perfect health to the eighth. Then the fever begins to seize them, and they keep their beds two days, very seldom three. They have very rarely above twenty or thirty in their faces, which never mark; and in eight days time they are as well as before their illness. Where they are wounded, there remain running sores during the distemper, which I don't doubt is a great relief to it. Every year thousands undergo this operation; and the French ambassador says pleasantly, that they take the small-pox here by way of diversion, as they take the waters in other countries. There is no example of any one that has died in it; and you may believe I am well satisfied of the safety of this experiment, since I intend to try it on my dear little son. I am patriot enough to take pains to bring this useful invention into fashion in England; and I should not fail to write to some of our doctors very particularly about it, if I knew any one of them that I thought had virtue enough to destroy such a considerable branch of their revenue, for the good of mankind. But that distemper is too beneficial to them, not to expose to all their resentment the hardy wight (sic) that should undertake to put an end to it. Perhaps, if I live to return, I may, however, have courage to war with them. Upon this occasion, admire the heroism in the heart of Your friend, &c. &c.
lw
There is a set of old women...
Set theory is getting to be over used.
In Seville, the old woman inoculates everyone who doesn't inoculate herself.
Hey Mossy. I would appreciate it if you stopped being an asshole to me. Could you please not address me at all? Thanks.
I would appreciate it if the commenter I'm no longer addressing stopped talking life-endangering bullshit based on nothing but their own wishful thinking.
Mossy,, you're being unreasonable for once. First, nothing said here endangers anyone's life; the world does not hang on the opinions of pseudonymous commenters on an eclectic web magazine. More seriously, a ban on bullshit would amount to a total lockdown of the economy of unfogged. A few million deaths is a small price to pay to avoid such a fate.
"I wish I'd had a brother instead of two whiny little kiddy sisters," said Peter. This was always recognised as indicating the high-water mark of Peter's rage. Bobbie made the reply she always made to it. "I can't think why little boys were ever invented," and just as she said it she looked up, and saw the three long windows of Mother's workshop flashing in the red rays of the sun. The sight brought back those words of praise: -- "You don't quarrel like you used to do." "Oh!" cried Bobbie, just as if she had been hit, or had caught her finger in a door, or had felt the hideous sharp beginnings of toothache. "What's the matter?" said Phyllis. Bobbie wanted to say: "Don't let's quarrel. Mother hates it so," but though she tried hard, she couldn't. Peter was looking too disagreeable and insulting.</The Railway Children>
If I see a friend and visit with them for an hour in a park standing 8 ft away is that okay?
My city's chief medical officer says that's okay, so long as you remain at least 2 metres (6.5 feet) apart:
During a teleconference on Wednesday, Vera Etches, medical officer of health for Ottawa Public Health, said talking to a neighbour over a fence is indeed ok, if the parties stay two metres apart.
96: I'll concede on the "life-endangering",* retract 87.2 as unnecessary, and hold that Megan is talking shit rather than bullshit.
I don't think unfogged runs on bullshit. I think it runs on irony: roughly, conscious stating of untruth on the assumption that the untruth will be transparent to interlocutors, highlighting the gap between statement and reality being the point of the exchange. The ironist knows that they're talking shit, and knows everyone else knows. Whereas bullshit is the stating of untruth in willful indifference to, or at least ignorance of, reality.**
I see no reason to think Megan knows she's talking shit, or is assuming everyone else can see she's talking shit. Whether she is indifferent to or merely ignorant of reality I don't know, but she in particular has earned no benefit of the doubt from me.
*Though Megan is presumably saying similar things in other circles where she may be listened to.
**I think Frankfurter settled on the former on the grounds that 'bullshit' implies bad faith; in fairness I don't don't think Megan is speaking in bad faith in this case.
95: I understand we're all under a lot of stress, but what the hell, man? Take it down a notch.
I've been thinking that given that COVID 19 is fairly safe for the under-40s, that it is probably worth doing variolation.
Based on studies in Wuhan, Case fatality rate for the 20-49s is 0.32%, and overall CFR for all ages is 2.3% but CFR for patients with no comorbidities is 0.9% - so, at a guess, CFR for healthy 20-49s should be 0.12% or thereabouts.
But I don't see how it could scale, so maybe it isn't worth it
Well, the relevant share of the US population is about 125 million. (That's 25-54, not 20-49, but it's probably not far off.) About half of those probably have cardio conditions or obesity or something like that, so they'd be excluded.
I think it would scale to kill about sixty-five thousand people.
(And once you'd done that you would still have a population that was mostly - 80% or so - susceptible.)
Not dissimilarly to Megan, I've been wondering what the ideal societal response would be under the following circumstances:
- Suppose we had infinite hospital, doctor, and nurse capacity, and a really crackerjack health care system.
- Suppose we knew for sure there would be no vaccine for years, nor any meaningful treatment besides ventilators and support.
How would we approximate life as usual without subjecting vulnerable people to constant fear of mass levels of dying and death?
The best I came up with was something like: People get classified according to their distance from vulnerable people. Are you one? Are you one degree removed? Two degrees removed? and perhaps the classifications max out at 3 degrees.
If you want to, say, visit your mother in a nursing home, you must maintain one-degree precautions for at least two weeks. Social safety nets would need to be robust enough to support people living at zero-degree and one-degree levels of restriction.
If you are two degrees and you're going to come into contact with someone who is one degree, maybe you have to meet in certain locations, wearing gloves and masks.
It kind of falls apart when I try to think about how to get the kids out of the house again. If schools open up, then it all falls apart. If schools don't open up, life has not been resumed in any meaningful sense (for me).
101: I think under those circumstances you do something like this:
First, you introduce social distancing until the case count drops way down - and get ready to put it back in as soon as case count goes up again.
Second, you introduce regular testing for as much of the population as you possibly can. Everyone who comes through an airport gets tested (having a TSA agent shove a foot-long swab up your nose wouldn't even be in the top five unpleasant things to happen to you at JFK). People in high-contact jobs like schoolteachers and so on get tested frequently. Medical staff get tested frequently. Everyone coming into contact with vulnerable people gets tested; if you want to see your mother in a care home, you have to get tested first, and all the care home staff are getting weekly tests.
Third, everyone who tests positive gets quarantined, and so does their entire household. Then you do massive contact tracing and test everyone they met in the last week.
Fourth, you get antibody tests working and start classifying people accordingly.
For that you don't even need infinite hospital capacity; you need large testing capacity.
I know mass testing is key, and that it seems to be working very well in South Korea and other places, but I don't understand how to square that with the idea that only 10% of tests administered to symptomatic people in Heebie County are positive.
The thing I want to know is what kind of rates of positive tests:symptomatic people are occurring in a place like Korea? I want to know if tests are being administered really poorly here, or what.
In Texas, people are too shy about ramming a swab up someone's nose before marriage.
I think "approximating regular life" would need to do something about in-home transmissions, which seem to comprise most of the spread. Quarantining the ill rather than their family, which mostly guarantees spread to the family.
Two lower income cities that are hotspots here joined together to open a motel with 140 rooms as a quarantine/isolation spot.
I'm wondering how contact tracing can work over the relevant time period (days) when mass transit is involved, and what that's going to mean for those of us who rely on mass transit to get to our jobs (I could maybe shift to bicycling, some part of the year, but my wife pretty much has to take a train to her office).
I've thought about that too. Check phone data to determine which exact trains/buses you were on, and do some privacy-invading phone company voodoo to text everyone on the same train?
I am also thinking that a decent person would be whipping myself into biking shape so I could keep off the subway when New York reopens. I've actually been running a fair amount, but not biking, and my 28-mile round trip commute is going to leave me super sore if I don't ramp up.
I regret finally giving away the bike trainer last year, after having and underutilizing it for a decade. I wonder if they're all on back order now.
102 is like watching newsom's press conference, with the supra depressing provisos that 1) i'm pretty firmly convinced that if california could ramp up sufficient testing capacity, the federal government would find some way to destroy-expropriate-forbid that capacity, and 2) it will all be for naught unless we can enforce quarantine on those travelling into california from other u.s. states, pretty confident again that the wretched fed judiciary we've allowed the gop to install will kibosh that as well.
so we are screwed pre-vaccine.
108: Yes. ROK, ROC, SG, PRC all doing this. It works.
101: the rules of conditional probability suggest the risk should drop quite quickly. If the probability of Alice transmitting it to Bob is 0.8, and Bob to Charlie is also 0.8, the probability of Charlie contracting it from Alice via Bob is 0.8*0.8 = 0.64, but once you get to Denise, it's down to evens.
This assumes though that Bob is symptomatic. If he's only a suspicious contact, the probability of secondary transmission will obviously be much lower and the potential chain peters out faster. (I was just trying to link this to R0 via Dunbar's number but remembered I don't actually know anything about epidemiology.)
I don't *always* have my cell phone on me. I worry a tiny bit about getting in trouble for not carrying one. I even have a friend with a flip phone who doesn't text.
109: 28 miles! I had no idea bike rides of that length were possible around your city. My commute is/was 17 miles round trip, I think, and I can feel my legs turning into lumps of inert protoplasm.
I liked this essay saying, among other things, the reason people fight contagious diseases instead of all the other things that kill people in equal number is that they love the opportunity to use techniques based on domination and control.
Mostly I'm just terrified on my mom dying alone because we're not allowed to visit until it's too late. I don't really think this is much of a situation with domination and control. The lockdown is basically toothless, tentative, and reactive here.
It isn't the force of the state that keeps me out of the nursing home. It's the knowledge that if I got the disease while traveling there, I'd be shoving a dozen families into the situation I fear.
118: Thanks. So far they are doing well.
https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=cordon+sanitaire+colonial+africa&btnG=
Here's what the OP author has to say about R and herd immunity: https://nucleardiner.wordpress.com/2020/04/10/the-two-numbers-i-am-watching-before-things-open-up/amp/
And that we're fucked without a vaccine.
123: makes sense. (Though why a chemist with an amateur interest in nuclear arms control is now blogging about epidemiology...)
re: 107/108
That is basically what they do in China. Everyone scans QR codes on entry to train carriages, buses, etc. So they know not only who has the disease but precisely where they were and with whom.
re: 113
I think it came up here before, but Dunbar was on some TV thing a while back -- 10 years ago, maybe -- and everyone (including me) was shocked that he's not some 100 year old emeritus professor (or dead for 40 years).
127: not difficult to do that in London, either. On a bus everyone scans in with an Oyster card or RF payment card. On a Tube, bit more difficult, but it should still be easy to say that they got on at X and got off at Y, therefore probably took this train on this line... though admittedly you couldn't narrow it down to a single carriage. From the payment card data you can normally get a phone number. And it need not involve TFL getting the phone number info - they just need to go to the banks and say "your customer with this number debit card was exposed today. Text them this warning please". Do the same thing for supermarkets given that most people who go into a supermarket buy something with a card just before they leave.
129: They experimented with some sort of wifi based tracking solution on the tube, didn't they? This might be the time to expand that. (Wildly optimistic take: they could then expand that into providing wifi access. Or that might be the carrot to get people more on board with it.)
While the UK is clearly not East Asia, it's interesting contrasting it to the US. People are genuinely willing to give up real or perceived individual liberties more easily here, if it is perceived to benefit the common good. No surprise there, but just interesting to see in action.
126 Isn't she a nuclear chemist who worked on denuclearization of some of the Soviet republics? I think she has some long built up expertise there, enough that Jeffrey Lewis takes her seriously. The epidemiology point is well taken though.
Wildly optimistic take: they could then expand that into providing wifi access.
They already do that. In stations only, but I that's what the journey tracking info was based on.
Dunbar is fun and very civilised: has, or had, a lovely anecdote about how he got his students to study the behaviour of people on the trains from Liverpool to London with their mobile phones (back when you had to qualify "phone" with "mobile"): if there was any group of men, they would always after a while put theirs on the table to see, at first, whose was largest; then the competition was to see whose was smallest. Now I suppose that it's who has the biggest screen again. But the way in which this competition for status was completely predictable delighted him.
132: Oh, that makes sense. I vaguely recall getting connection in some of the deeper stations. But ideally they'd expand it to in-carriage trackers for appropriate granularity and Big Brotherness.
Once we get to the point where purposeless travel doesn't feel like attempted gerontocide, I'm going to have to take some aimless tube journeys. Still haven't gotten a chance to go to that mysterious place at the far end of the line, Cockfosters.
FWIW, I'm not sure Oyster card tracking would work even as a stopgap. I'm pretty sure you can still get one without registering it to an identity.
135: maybe - I don't know - but even if you can get an anonymous Oyster card, how do you top it up? By transferring money on to it from a debit or credit card. And there's your link - I seriously doubt there are many people out there who have anonymous Oyster cards and only ever top them up with cash.
I dunno. It used to be a thing that lots of corner shops did. Maybe still do, the ones that haven't become vape shops.
I'm also not sure if the system currently ties the payment ID to the Oyster ID beyond the immediate transaction unless you set up recurring payment, though obviously that could be fixed with some programming work.
131. Speaking of Soviet republics, I hadn't realised that Lukashenko is pretending COVID-19 doesn't exist. The Belarusian health service is trying to crowd fund basic equipment.
135: I think that's not a problem, if it's not a big segment of the population. If you're tracking people for terrorism or crime control, you care if there's a practical way to evade it even if it's uncommon behavior, because the criminals are the ones who are going to intentionally evade surveillance. But for disease control, the occasional anonymous weirdo isn't specifically more likely to be infectious, they're just a slight imperfection in the system. It's not ideal, but it doesn't wreck everything.
I have an anonymous Oyster card bought from a ticket machine, which I've only ever topped up at ticket machines (admittedly with a debit card, but I bet you can do it with cash). But I very rarely use(d) London Transport - going into central London I'd normally either walk or cycle from Kings Cross.
I think the term for having an anonymous Oyster card should be."oyster cracker."
Someone who has an anonymous Oyster card to conceal their identity as they do crimes uses it as an Oyster shell.
I have an anonymous Oyster card that I've probably topped up with a credit card before.
But for disease control, the occasional anonymous weirdo isn't specifically more likely to be infectious, they're just a slight imperfection in the system.
I don't think it's just anonymous weirdos though. It's pretty much all tourists (including intra-UK tourists).
I don't think there are going to be many tourists in London for a wee while yet.
Nonsense! There'll be millions, just as soon as you've finished up those FTAs with the US and EU.
144: Even tourists are mostly going to use credit or debit cards, like Ume.
144 is right, but, I think, probably still a relatively small percentage of London travellers. That said, maybe a more disease-vector-y than normal segment of them. Especially on the Central and Piccadilly lines, and Gatwick and Heathrow trains.
Germans don't use credit cards? I'm not doubting you, it's just a fact? Stereotype? that I was completely unaware of.
150: Germany is (somewhat notoriously) a much more cash-run society than other comparable European countries. Two cultural attributes overlap. First is a desire for privacy shaped by experience with two different dictatorships. Second is a price sensitivity that is greater than other European societies. Many small German business won't take electronic payments *at all* because they are unwilling to pay the fees that the providers charge and because they are unable to pass those costs along to their customers. Some will take *this* kind but not *that* kind. My local Indian place will take an EC-branded debit card but not a Visa-branded debit card, nevermind that making it difficult for customers to give you money may affect whether people are willing to walk in your door at all.
All of this is domestic, though. I would be very surprised if German tourists in London didn't use electronic payments at some point in their trip. On the other hand, while visiting London from Berlin, I did indeed purchase and later top up my anonymous Oyster card with cash. And on the other other hand, German visitors to other eurozone countries might well continue with the cash-y ways they are accustomed to at home.
Largest single day increase in cases here. +560. Fuck.
Lots of small businesses in Pittsburgh take only cash for similar reasons. Possibly also to make it easier to pay less taxes.
Almost nobody in the UK takes American Excess any more, because their terms are such rubbish.
Good takedown of the IMHE model here.
Do you have to physically press the Oyster card against its reader, or can you hold it a centimeter away? I did the slap-and-hold thing when I was in London, which makes finger-contact likely, but I think I just assumed you needed to.
Spoke too soon about cases leveling off low. After a fews days with five and six, yesterday we had eleven. Today seven. That's statewide. We're at 33 here. Rural Toole county is at 29 -- 24 of those are women, and 9 are women 80+. One old folks home up there accounts for it.
Two term former Republican gov Racicot put out a video thanking everyone for social distancing, and urging everyone to keep it up. On a different timeline, he ends up as GWB's VP and we get a Republicanism that still favors money over people, but on class lines, and without all the science-denial Confederacy worship stuff we've gotten in this one. I'm sure the prospect of right wing nuttery motivated Racicot to action.
I know about the Racicot video because Gov Bullock retweeted it, with thanks.
156: It's a Philips/NXP MiFARE RFID chip so in principle shouldn't need actual contact, because radio, but in practice you'll be lucky to get through the gateline without slapping it down on the reader. That said there doesn't seem to be much or any evidence of people catching it from surfaces.
If you do phone payment you can hold it further away.
The buses are still running here but they have blocked off the front of the bus to protect the driver. I assume that means there's no way to collect a fare, but maybe something was improvised.
161: They did something like that for a while here, and it did indeed involve waiving all fare collection. They eventually decided it was still too much risk and shut down the bus system entirely. Essential trips can still be done but they're on-request, point-to-point deals with the smaller paratransit van-like buses.
A headline so good you don't need to bother with the article: https://nymag.com/intelligencer/2020/04/trump-coronavirus-liberate-michigan-minnesota-virginia-lockdowns-tests.html