Stay in your own lane?

Even though there are scientists who have the kind of expertise that might help us to better understand this pandemic, there’s a tendency to suggest that it would probably be best if they stayed in their own lane. Although I do have some sympathy for this, I think it’s too simplistic; researchers should be free to study what they wish, and those within a discipline should be willing to listen to people with relevant expertise from outside their discipline. However, there are certainly examples when researchers have tackled problems outside their core area and not made particularly constructive contributions. The problem, though, isn’t that people don’t stay in their lane, it’s that they don’t do their homework properly when they move outside their lane.

One of the most difficult things about doing research isn’t the technical aspects, it’s being very familiar with the details of a topic, and knowing what questions are worth asking. Just taking some data and throwing some analysis method at it isn’t very useful if you don’t understand how the data was collected, it’s limitations, or the significance of the analysis in this particular context. For example, the impact of this virus is almost certainly going to depend on the strategy that is employed. Hence, you can’t really infer anything about an alaysis if you don’t take into account what strategy has already been employed and how this strategy might evolve. There are, of course, many other factors that should also be considered; it’s clearly not simple.

What motivated this post was a recent post by Andrew Gelman that highlights how to be curious instead of contrarian about COVID-19. It was itself motivated by an article by Rex Douglass that provided Eight Data Science Lessons, using an article written by a rather contrarian, and not very curious, lawyer to illustrate what not to do.

In my view, the key thing is that even though these are unprecedented times, it doesn’t mean that we should take research short-cuts. As the articles above highlight, we should be familiar with the topic, care about the research questions, be careful about the design of the research project, be willing to revise our understanding if the model doesn’t match the data, or if new data becomes available, and be very clear about assumptions, uncertainties, and the overall context. I would add that we should also be willing to “trust” other experts. If we want to live in a world where people listen to us when our expertise is relevant, we should be willing to listen to other experts when their expertise is relevant.

This entry was posted in Contrarian Matrix, Philosophy for Bloggers, Scientists, The philosophy of science, The scientific method and tagged , , , , , . Bookmark the permalink.

250 Responses to Stay in your own lane?

  1. John Hartz says:

    There’s been a lot of chatter in the US media about the “models” being used to forecast the spread and severity of COVID- 19 — particularly one developed by the University of Washington. Does anyone happen to know whether these models are run on supercomputers? I also wonder if some of the people responsible for the care and feeding of GCMs are now temporarily working on modeling the pandemic.

  2. jamesannan says:

    Treading on the toes of giants rarely goes well. IME. Being right probably makes it worse.

  3. jamesannan says:

    I think even the state of the art models are computationally reasonably cheap. They don’t have anything like the size/”resolution” common in GCMs. I might be wrong. One was specifically described as 13,000 lines of code and sounds like it was mostly written by one person or perhaps a small group.

    Of course you can make anything expensive with a large enough ensemble…

  4. James,

    Treading on the toes of giants rarely goes well. IME. Being right probably makes it worse.

    Yes, there is that. There are, of course, examples of people straying out of their lane and making very substantive, and constructive, contributions 🙂

    I think even the state of the art models are computationally reasonably cheap. They don’t have anything like the size/”resolution” common in GCMs. I might be wrong. One was specifically described as 13,000 lines of code and sounds like it was mostly written by one person or perhaps a small group.

    Yes, that’s my understanding too. These are not computationally-expensive computer codes, as far as I’m aware.

  5. dikranmarsupial says:

    “Adding people to a late software project makes it later.”

    [Fred Brooks – “The Mythical Man Month”]

    There is nothing wrong with wanting to experiment with data and see what you can do with your favourite model*, but we are running late in our response to COVID-19, so unless you have good reason to think you really have something to contribute, please don’t get in the way of those who actually know what they are talking about (and ideally don’t spread misinformation either).

    * “Statisticians, like artists, have the bad habit of falling in love with their models. ”

    [GEP Box]

  6. James,
    If I understand your general criticism, it’s that they’ve underestimated the impact of the virus (as measured by something like R0) even though the data was clearly indicating a doubling timescale that was considerably shorter than would be suggested by their parameters. Is that about right?

    I must admit that I’m slightly concerned about the Imperial College code not being public. I tend to think that either codes should be publicly available, or it should be clear how to develop an equivalent code. Given the timescales involved here, though, the latter is probably not realistic, and so their code really should be public – IMO.

  7. jamesannan says:

    Yes ATTP that was the most immediate and obvious problem. There was ample data available at the time to show that the doubling time was quicker than they were modelling, and it is well known in that community that parameters (especially R0, but maybe some others too) are context-dependent. Their parameters choices come from an analysis of a small data set of infections at the start of the Wuhan outbreak. It is really baffling to me that they made no attempt to fit the trend in the data. A large R0 has important implications for the possibility of suppression and also the timing and magnitude of the peak so this isn’t a trivial or insignificant detail.

  8. jamesannan says:

    They also did a very weird and IMO pretty shady thing in the 16 March report where they used one value of R0 to show that mitigation would lead to a lot of deaths and then used a smaller value to show the potential for suppression!

  9. James,
    Thanks. Yes, what you describe in your latter comment is rather odd.

  10. BBD says:

    I had to smile – thinly – at the advice given in the linked Douglass essay on How to be curious instead of a contrarian about COVID-19:

    Don’t Use Straw Men to Represent the State of the Art.
    Don’t claim the state of the art ignores factors that it actually takes into account.
    Don’t get basic facts wrong.
    Don’t cherry pick data.
    Take the consensus explanation seriously.

    It seems Epstein is treading a well-worn path.

  11. anoilman says:

    Unlike global warming deniers its normal for anyone technically inclined to examine and question what’s out there. But its a big leap to make it from that, to, actually solving wicked problems.

  12. Dan Riley says:

    Most of the epidemiological models are simple compartmental models like SEIR, which are rather like box models in climatology. Simple, fast, make a lot of assumptions about homogeneity, etc. However, there are some groups doing agent-based models that simulate individual people and interactions; these are closer to GCMs in scope and detail, and take substantial amount of computing power for anything much more ambitious than a cruise ship population. See for example

    https://www.statnews.com/2020/02/14/disease-modelers-see-future-of-covid-19/

  13. Dan,
    Thanks. Yes, I realise that some of models have quite a lot of detail. I do wonder, though, if even these are quite as computationally expensive as some of the big GCMs, for example.

  14. Dan Riley says:

    AIUI, GCMs can eat whatever computational power is available just by making the grid finer. I would think that the corresponding improvement for an agent-based epidemiological model would require finer-grained behavioral data as input. Naively it seems at least plausible that would limit the computational complexity, but I don’t really know.

  15. Dan, yes, that’s broadly my understanding too. I’m not an expert, though, so am also not sure.

  16. Steven Mosher says:

    “AIUI, GCMs can eat whatever computational power is available just by making the grid finer. I would think that the corresponding improvement for an agent-based epidemiological model would require finer-grained behavioral data as input. Naively it seems at least plausible that would limit the computational complexity, but I don’t really know.”

    I suppose if you had a model with 7 billion agents, roads, buildings, trains, planes,,
    door knobs, areosols, etc, behavior likelihoods, detailed demographics with co comorbidities…..
    There is no bottom.

    Hmm. searching for an analogy.

    In war simulation one of the things we struggled with was the probability of kill for a missile
    versus a target. at one end you had guys who wanted detailed models. CAD model of the aircraft
    CAD of the missile, detailed physics of the blast, fragments, what missile parts hit what critical
    systems. To get test data for these guys they would do stuff like take 20 airplanes and
    shoot them up in the desert. So, detailed detailed physics. There was NO BOTTOM in this
    approach. No level of detail would satisfy some types of minds. At the other end
    were the paramatizers. The missile is a black box. If I deploy my box against you, you die 50%
    of the time. Like Dungeons and Dragons type approach. No physics needed I shoot you
    roll the dice, etc. Don’t like 50%/ ok we try numbers from 25% to 75%. Here’s my ensemble.
    I win for all reasonable values in the box.

    The “effects based” approach of weapon versus target gets codified into a nice little book
    called JMEM, it literally is like D&D. look up a weapon ( Mk-82), look up a target (bridge)
    read the PK, look at the CEP to get an idea of the blast area (CEP is circular error probability)
    drop 5 bomb, bridge destroyed. Your move.

    This worked for getting the actual job done of destroying stuff, but it DID not work for things
    like collateral damage or dropping bombs in civilian areas. So a detailed model and
    lawyers were required. and a program called Bugsplat.

    https://www.washingtonpost.com/archive/politics/2003/02/21/military-turns-to-software-to-cut-civilian-casualties/af3e06a3-e2b2-4258-b511-31a3425bde31/

    “Approved just two months ago, the program represents a significant departure from the traditional method of drawing a simple circle around a target to show a bomb’s estimated blast effect and determine what civilians might be at risk nearby, Air Force officials said.

    Instead, Bugsplat generates blob-like images — resembling squashed insects — that military officials say more precisely model potential damage by a particular type and size of bomb dropped by a particular aircraft flying at a given altitude. This enables commanders to fine-tune attacks and, in some instances, can embolden them to order bigger bombs than they would have employed relying on less sophisticated modeling methods, the officials said.”

    https://muse.jhu.edu/article/735792

    As with all things modelling there was this tension between the “effects level” guys and the more
    physics based, engineering type approaches. Mo detail please!. But there literally is no bottom
    and no way to do experiments to let you know what level of detail was sufficient.. what detailed physics process could be parameterized and replaced with an “effects” level model.
    ( hey we dont have good data on bombs versus hospitals, lets drop a few and collect some data!)

    And so decisions came back to “When do you need the answer ?”
    Oh, you need to decide now? effects level for you.

  17. Below is a link to the Supplementary material for this paper. It describes three different models that all considered the same region. They do consider households, work, schools, travel, etc. They use census data to determine how people are distributed, and also things like age distributions. I wasn’t quite sure if they modelled all the people or if they typically used 2000 person groups to represent some larger number of people.

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2290797/bin/pnas_0706849105_index.html#ST

  18. Dave_Geologist says:

    I posted this at the end of the previous thread but it’s relevant here. The agent-based models are computationally expensive (but the hardest part seems to be gather population data so they’re not GIGO), so they ran analytical models to do sensitivity tests on various measures for a fast turnaround. Checking them in the base case against the agent-based model. But I can’t imagine they’re nearly as expensive as a modern GCM. The Imperial agent-based one was published about fifteen years ago (ATTP’s link is to a later iteration), and probably developed and run on the technology of twenty years ago.

    A very good summary of the Ferguson models is on Nature. Models plural: agent-based, differential-equation based, even some Bayesian; what changed in the UK decision-making (China data indicated half of ICU patients could get by on positive-pressure oxygen but Italy showed most need intubation); European R factors were looking higher than China as they came in; other teams were consulted; capacity to expand ICU capacity was initially over-optimistic; etc.

    IOW they started by not by using made-up numbers but by using China data because it was all they had; then added Italy data as they were the European canary; then added more European and UK data as it came in; are continually updating; it’s not a single model; and it does have uncertainty ranges. As Keynes (allegedly) said: “when the facts change I change my mind: what do you do?”.

    Most of the new data is not yet peer reviewed, but a panel of dozens of epidemiologists and clinicians is the best peer review team there is, worth 20 pairs of journal reviewers. They’re well able to sort the wheat from the chaff, for all but data fraud which peer review doesn’t pick up anyway.

    Ferguson is polite and cagey but

    Ferguson says the significance of the model update might have been exaggerated. Even before that, he says, models already indicated that COVID-19, if left entirely unmitigated, could kill in the order of half a million UK citizens over the next year and that ICUs would be stretched beyond capacity. Advisory teams had discussed suppressing the pandemic by social distancing, but officials were worried that this would only lead to a bigger second outbreak later in the year. Widespread testing of the kind seen in South Korea was not considered; but, in part, says Ferguson, this was because Britain’s health agency had told government advisers that it would not be able to scale up testing fast enough.

    As for the Chinese data on ICUs, clinicians had looked at them, but noted that only half the cases seemed to need invasive mechanical ventilators; the others were given pressurized oxygen, so might not need an ICU bed. On the basis of this and their experience with viral pneumonia, clinicians had advised modellers that 15% was a better assumption.

    The key update came the week before Ferguson briefed government officials at Downing Street. Clinicians who had been talking to horrified colleagues in Italy said that pressurized oxygen wasn’t working well and that all 30% of the severe hospitalized cases would need invasive ventilation in an ICU. Ferguson says the updated models’ mortality projections didn’t change hugely, because many predicted deaths are likely to occur in the community rather than in hospitals. But the understanding of how health services would be overwhelmed, and the experience of Italy, led to a “sudden focusing of minds”, he says: government officials swiftly pivoted to social-distancing measures.

    rather suggests that the herd immunity approach had been seriously considered and it was only when the full scale of what it implied confronted minsters and advisers that they had their “oh shit” moment. Footage of people dying when they could have been saved with more ICU capacity has a way of focusing the mind.

  19. Dave,
    Thanks, that’s is a good article.

  20. Chubbs says:

    Utube interview from Professor Ferguson on Feb 5. Covering uncertainty, projections, implications etc. Interesting to see how much of the uncertainty has been resolved. At the time I found it highly informative and concerning. The real value is not the model itself, but understanding the limitations, and the implications from the key parameters. Turns out he should have been even scarier.

  21. Dave_Geologist says:

    And on open-sourcing. Note how quickly the thread gets swamped by idiots, just like a climate-modelling thread. Some of them the self-same idiots, I bet. At least if it’s on Github serious, good-faith actors with something to contribute will be able to access the real deal and feed back improvements for evaluation and possible incorporation. Of course that will be too late to inform this round of decisions. And he’ll be able to palm off vexatious, time-wasting requests from people who just want to restart the economy and win Trump a second term. Hmm, that last part sounds familiar…

    As Ferguson said in the Nature article, he’s been working on this non-stop since January (including, but he’s too modest to make much of it, when he had Covid-19). I’m sure he has more important things to do than satisfy the whims of armchair epidemiologists. They can use Excel from their armchairs. Hey, I’ve done it myself: fit to the exponential total-deaths stage, multiply by 100 for the number of cases, and extrapolate it to 60-80% of the population infected to get the absent-social-distancing total deaths two weeks later assuming an infinite supply of ventilators and ICU beds. Tweak the death rate up or down as you please to change the asymptomatic numbers, but unless it’s off by orders of magnitude you’re looking at many hundreds of thousands of UK deaths and millions of US deaths. As someone commented on the Oxford study, a hundred or a thousand times as many asymptomatic cases would be proved or disproved in Italy within the week, and had probably already been disproved in Lombardy at the time of publication. Add on the people left on trolleys when the ventilators run out. You don’t need to know whether it’s 250k, 500k, 700k or 900k in the UK to make a yes/no decision on lockdown. Or indeed to know that Britain duplicating Lombardy a few weeks later when we had all the warning we needed would be political poison for whoever was in charge. Just as you don’t need to know whether ECS is 1.5K, 3K or 4.5K to know that climate action is required.

  22. Dave_Geologist says:

    Oops, got so carried away there I forgot the Twitter link 😦

    And be safe, everyone.

  23. izen says:

    @-Dave_G
    “Footage of people dying when they could have been saved with more ICU capacity has a way of focusing the mind.”

    The ‘no win’ problem for advanced societies with a good healthcare system is that ICU capacity is some-what irrelevant.
    Treatment on a ventilator is not very effective, the death rate varies from 60% for the fit young to 85% for the old, ill and obese.

    This makes the whole drama about how many ICU and ventilator beds are available something of a PR distraction. As other Nations have demonstrated the most effective way to deal with this is to test for infective carriers, trace all possible at-risk contacts and isolate them all until ‘Nature’ takes its course.
    The focus on hospital treatment for the % who develop serious respiratory failure is a sideshow in terms of controlling the amount of deaths this pandemic will cause.

    Reducing the transmission rate is the only effective response until either herd immunity increases to the point it is inherently rare and unlikely to be caught, or until an effective anti-viral or vaccine treatment is available. The ‘concern’ about ICU/ventilator availability is either an intentional, or expedient but unwitting distraction from the only effective method of control that we have available for this disease at present.

  24. Dave_Geologist says:

    Turns out he should have been even scarier.

    Chubbs, I’m reminded of an interview years ago with a current or former government scientific adviser who said something along the lines of (my metaphor) “sometimes you have to accept half a loaf in your formal advice, because if you insist on a whole loaf you’ll just be ignored and there will be no loaf. You’ve mitigated as best you can, and can always try to incrementally improve later.”

    Fauci and Birx are no doubt having to navigate that minefield:

    ‘I’m going to keep pushing.’ Anthony Fauci tries to make the White House listen to facts of the pandemic

  25. Dave_Geologist says:

    On a lighter note (o/t but I’ll claim a tenuous link to exoplanets 😉 ): Superconductivity discovered in extra-terrestrial objects for the first time

    During their search, Wampler and colleagues have discovered superconducting alloys comprising indium, lead and tin in two very different meteorites.

    Pah! Obvious, innit? Borg solder for superconducting circuitry. There must be a wrecked Cube out in the Kuiper Belt.

  26. Dave, Many elemental metals show superconductivity.

  27. Chubbs says:

    “Faux and Birx are no doubt having to navigate that minefield”

    Any expertise is heavily discounted in the pandumbic:

    https://mobile.twitter.com/TheDailyShow/status/1246146713523453957

  28. Dave_Geologist says:

    On the “do you need a supercomputer” question: no.

    There’s a piece about the model (which from the open-sourcing thread seems not to have been greatly updated in terms of code) from 2005 (for Thailand, not the UK or US).

    This meant running the model hundreds of thousands of times. To do these runs quickly, the model needed to be coded efficiently, and required computers with huge amounts of memory — 20 times that found on a typical PC. In fact, the team hooked up ten high-powered computers in parallel, but even then the final runs took more than a month of computer time.

    One of the advantages of building your own computers is that you can check your emails for the hardware you bought. The most the motherboard I bought in 2005 supported was a Pentium 4 Extreme Edition 3.2 GHz CPU, single-core but double threaded. 32-bit of course so let’s assume max 4Gb memory but most people, me included, will have had 2Gb (I bought two 512Mb sticks later the same year, probably to fill the two unused slots). So they were running on ten workstations with 400 or 800Gb memory in total. I had just switched at work from an SGI Octane running IRIX (but to run big models we used a shared Indigo in the corner which blew papers off desks 20 feet away when the going got tough and the booster fan kicked in…) to a Sun Ultra running Red Hat. It was night and day, and same again with the next upgrade to an HP which IIRC had 64Gb memory and 16 cores. So let’s assume a brand-new high-spec Ultra, 32Gb memory and 4 cores. Ignoring clock rate because since 3GHz it’s all been about cores not speed, Thailand took a month with 300-400Gb and 40 cores. A single HP Z8 under-desk workstation can have up to 3Tb and 56 cores. It’s powered by a 1,700-watt supply and cooled using invective ducting, so it won’t even blow papers off the desk 🙂 . Ten of them should run Thailand in a day or two. And that’s before you even start talking mainframes.

  29. verytallguy says:

    The Rex Douglass article is really good, thanks for sharing.

    It’s depressing to see how much air time these bullshitters get and how many people are convinced by their self assured rhetoric.

  30. Chubbs says:

    This didn’t age well:

    “If we assume that case fatality rate among individuals infected by SARS-CoV-2 is 0.3% in the general population — a mid-range guess from my Diamond Princess analysis — and that 1% of the U.S. population gets infected (about 3.3 million people), this would translate to about 10,000 deaths. This sounds like a huge number, but it is buried within the noise of the estimate of deaths from “influenza-like illness.”

    https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/

  31. Joshua says:

    Stop with the Diamond Princess.

    Results from an SES outlier cohort, subjected to non-typical conditions, resulting from a potentially anomolous strain, should not be considered generalizable.

  32. JCH says:

    Study this group of ~40,000 people:

    https://i.imgur.com/ZbkQhHG.png:

  33. John Hartz says:

    The introductory chapters of a very interesting an informative article…

    An invisible enemy brings life to a halt. People become isolated and panic. Neighbors start seeing one another as threats.

    If life during the Covid-19 pandemic makes it seem like you’ve entered “The Twilight Zone,” that seminal sci-fi series about dread and paranoia, than you’re more right than you realize. On March 4, 1960, it aired a classic episode that’s a cautionary tale about how social order can quickly break down when an unseen threat causes fear to go viral.

    The episode, titled “The Monsters are Due on Maple Street,” examines what happens to a leafy suburban neighborhood when it’s suddenly hit with an unseen menace. The power goes out and rumors spread of an alien invasion. The residents of Maple Street suddenly see sides of their neighbors that shock them.

    Part of the genius of Rod Serling, the “Twilight Zone’s” creator, is that he depicts how people react to fear and paranoia in ways that remain timeless. But perhaps no episode of Serling’s celebrated show captures so well what so many Americans are experiencing now — and how grim life could quickly get if people aren’t careful.

    A classic ‘Twilight Zone’ episode warns us how not to behave during a pandemic by John Blake, CNN, Apr 4, 2020

  34. John Hartz says:

    Disclaimer: I’m old enough to have watched the original broadcasts of the Twilight Zone during my formative years. I was enthralled by them. I vaguely recall the episode, “The Monsters are Due on Maple Street,”and will watch it again later today.

  35. Steven Mosher says:

    “Stop with the Diamond Princess.

    Results from an SES outlier cohort, subjected to non-typical conditions, resulting from a potentially anomolous strain, should not be considered generalizable.”

    yup

    1 out of 51 80+ year olds died on the diamond princess.

    I mean FFS, if there is one bit of data I toss, it’s diamond princess.

    I looked for Diamond princess strains on nextstrain.org. Zip, none uploaded
    other cruise ship is uploaded.

    the view from experts is that the 8 strains are not that different, hence evidence they are not
    under selection stress. Means, they found a good host, easy to spread and doesn’t kill
    too many. Goldilocks!

  36. An_older_code says:

    Rex’s daming conclusion of Epstein psuedo scientific approach

    “I don’t care who’s a professional or not you don’t win an argument by showing you got a PhD you win an argument by entering into a public debate with people who disagree.”

    Reminded me of the approach Nick Naylor (the eponymous hero of “Thankyou for Smoking”) took when telling his son, when he said

    “I don’t have a MD or law degree. …”

    Before going on to explain

    “That’s the beauty of argument, if you argue correctly, you’re never wrong. …”

    Also another chance to repost this wonderful take down of the “i Think, and My Guess” brigade

  37. Ben McMillan says:

    I wouldn’t be so quick to dismiss the Diamond Princess data: despite wide error bars and the passengers being unrepresentative (SES), this data is useful because of the systematic testing of those aboard. It actually lines up pretty well with other data as long as you treat it as an order-of-magnitude estimate. Especially if you don’t take the already small number of cases and try to break it down by age (you need other data to correct for age profiles).

    Real epidemiologists have used this data to good effect…
    https://cmmid.github.io/topics/covid19/severity/diamond_cruise_cfr_estimates.html

    Basically, this data is enough to tell you that a full-blown epidemic is going to overwhelm your health system and lead to a large increase in overall mortality over a short timespan. It definitively rules out the idea that the vast majority of cases are asymptomatic. This is consistent with what happened in Lombardy.

    The people arguing that that the Diamond Princess data shows we have little to worry about were still predicting massive fatalities in their do-nothing scenario. Mostly (in addition to overconfident low-ball predictions) they were arguing that it didn’t really matter if a whole bunch of old people died over a short period of time.

  38. Dave_Geologist says:

    Chubbs, absent social distancing and lockdown, what miracle will stop the infection at 1% of the US population? Herd immunity won’t be achieved until 60% (more like 80%, using the latest European pre-lockdown R values which are about 50% higher than in China). The USA may be in between or more like China. Its doubling time for deaths is a tad longer than Europe’s, which suggests less interaction normally. Big country, cars for everything, less of the kissing and hugging and extended-family living which occurs in the Latin countries which dominate the European numbers. But China is nothing to be complacent about.

    Re the various Diamond Princess comments. The death rate on the Diamond Princess was half that for equivalent age cohorts in China. Estimating the infection and case fatality ratio for coronavirus disease (COVID-19) using age-adjusted data from the outbreak on the Diamond Princess cruise ship, February 2020 (I may have shared a preprint earlier: the numbers are slightly tweaked but the essential conclusions are the same). And of course all the serious people use age-stratified deaths to account for different population distributions in different situations or countries. Including those which calibrate to the Diamond Princess.

    Old, rich Westerners are, on average, healthier, better nourished, fitter etc. than old, poor Chinese retired farmers and factory workers. Who’da thunk it?

  39. Ben McMillan says:

    The take-home message is that any of the serious mortality estimates suggest that a full-blown unmitigated epidemic would be appalling. But hopefully this message is mostly getting through.

  40. Everett F Sargent says:

  41. Chubbs says:

    Dave, Joshua – Yes, as I posted above, the estimate from Mar 17 didn’t age well. US will blow by 10000 either today or tomorrow so estimate only took a little over 2 weeks to bust horribly. Interestingly the author is in his “lane”

    “John P.A. Ioannidis is professor of medicine and professor of epidemiology and population health, as well as professor by courtesy of biomedical data science at Stanford University School of Medicine, professor by courtesy of statistics at Stanford University School of Humanities and Sciences, and co-director of the Meta-Research Innovation Center at Stanford (METRICS) at Stanford University.”

  42. Joshua says:

    > because of the systematic testing of those aboard.

    That goes to one of my questions – which whether the treatment conditions could be considered generalizable.

    In the one hand, it seems a perfect environment for spread.

    Im the other hand, once countermeasures were launched, it’s is a perfect environment for isolation/social distancing.

    Chubbs –

    I have mad respect for Ionnidis. And yes, I should walk back my criticisms. He’s obviously more knowledgeable about this than I, byorders of of magnitude. Plus he’s no doubt smarter.

    I guess we can chalk my confidence up to DK.

    I should limit myself to saying that I don’t understand why people should consider the Princess Diamond data as generalizable.

  43. Joshua says:

    In contrast:

    > Our analysis shows the importance of adjusting for delays from confirmation to outcome in real-time estimates of fatality risk, and the benefits of combining datasets alongside appropriate age adjustments to provide early insights into COVID-19 severity.

    As someone not remotely expert, it seems to me that combining datasets is important, and extrapolating from single datasets with an array of unique attributes, not so much.

    But I do understand the importance of listening to people more knowledgeable than I.

  44. Chubbs says:

    Too many of these out of lane projections focus on the body count, an uncertain quantity. Use of medical resources is much more important from a public health or economic standpoint.

  45. Dave_Geologist says:

    I should limit myself to saying that I don’t understand why people should consider the Princess Diamond data as generalizable.

    Because they know what’s generalisable and what’s not. And that what’s generalisable for one purpose is not for another.

    For example, mortality rates are stratified by age group. One study notes that on that basis the mortality rate was half that in China. They show what China’s true mortality rate would be if it was actually the same, which increases the number of asymptomatic cases there by a factor of two or three. Possible, contacts were traced and tested, everyone with a fever was tested, but not everyone with no contacts and no fever was tested, only those arriving in other cities’ stations and airports. But others ignore the ship and stick to China. Perhaps, like me, thinking that elderly cruise ship passengers are unrepresentative of other elderly people (more resilient, not less). Just as you get a range of ECS values, even excluding contrarians. Advisers read all the studies, weight them, and come to a consensus.

    From the cruise ship studies, they know the R factor when they were openly mingling and the R factor when confined to their cabins. Neither of those corresponds to R factors in the wild, in or out of lockdown, and neither is used to plug in wild-case R factors. OTOH the number of asymptomatic cases is a good harvesting ground because it’s like a Lensky experiment for sick humans, isolated and carefully monitored.

    The Chinese data suggested that more patients in ICU could get by without ventilators than turned out to be the case in Italy. Advisers then had to decide which to use. In the UK, they applied the precautionary principle and used Italy. I would have too. Eventually we’ll find out why there was a difference, but meantime we have to decide whether to play safe or gamble. Etc.

  46. Chubbs says:

    Nic Lewis has “discovered” that the odds of dying from covid-19 are about the same as the odds of dying in any one year, plus or minus an uncertainty range. If you are older, your odds go up rather dramatically, fancy that. A rather common mistake for someone out of their “lane” is overconfidence about a complex subject based on a single uncertain aspect. Sound familiar?

  47. Dave_Geologist says:

    Did Nic start with an objective Bayesian prior of zero deaths per year? In both categories?

  48. David B Benson says:

    For the USA:
    https://covid19.healthdata.org/projections

    From a professional modeler @ the University of Washington.

  49. Steven Mosher says:

    “In the one hand, it seems a perfect environment for spread.

    Im the other hand, once countermeasures were launched, it’s is a perfect environment for isolation/social distancing.”

    Yup. I never quite got the argument that it was the “perfect place” for spread.
    as if anyone knows what the most effective way of spreading it is.

    Other “perfect” environments

    1. The case of the BUS in Wuhan. 1 guy, spread to 6, one who got on the bus 30 minutes later
    2. the restaurant in Guangdong china, 6 people meet for dinner,
    3. Church service in Korea 1 lady, 1000 victims.
    4. Kirkland nursing home.. 1 person lead to dozens of infections.
    5. call center in Seoul. 1 person leads to over 100

    so what makes the diamond princess “perfect”? How often do they clean surfaces?
    what is the ventilation like? how many people do you actually interact with on a cruise
    ship.. are you walking around shaking hands with everyone? big elevators or small?
    how fastidious is the kitchen staff?

  50. Everett F Sargent says:


    The raw dailies suggest that most countries have peaked (assuming there are no so-called weekend effects in counts). Three countries still stand out as to current (small) doubling time and current (large) body count: FR, UK and US.

  51. Steven Mosher says:

    Patient 31

  52. Dave_Geologist says:

    We don’t have to guess about the Diamond Princess Steven, we know. The thing that’s perfect about it is that it was a closely monitored Petri dish. Transmission potential of the novel coronavirus (COVID-19) onboard the diamond Princess Cruises Ship, 2020. Not that it’s a perfect analogue for other settings.

    R up to 11 during the early, pre-quarantine phase. Less than 1 after cabin lockdown had done its job. Perfect for spreading (3 or 4 times China, 3 times Europe, similar to the much more infectious measles in the wild). That’s doesn’t tell us why it was perfect for spreading, but I can hazard a guess or three. Some of which will be replicable in western cities, others not (the Wuhan lockdown was like the quarantined cabins – no-one left home, all food delivered). Some of which will be relevant to settings like care homes but not to farmers’ fields or factories.

    The crew (Fig. 2C) are interesting. Small numbers so huge error bars, but they carried on through the lockdown as if nothing had happened. Which, other than more meticulous hygiene, they probably did in real life. Speculation now but there might be a lesson there for key workers who can’t do their job without breaking social distancing rules. It must have been impossible not too during much of their duties, even if they left the trays outside cabin doors. Their later drop-off too. Does that suggest some reverse contamination despite the precautions, until the community infection rate fell below a threshold?

    I recall some passengers saying they didn’t have windows so I presume they kept the a/c on or they’d have been broiled. But cruise ships may have more rigorous disinfection and filtering of a/c than most in light of past outbreaks of Legionnaire’s Disease etc., so that may not be generally applicable.

  53. Everett F Sargent says:

    ” The thing that’s perfect about it is that it was a closely monitored Petri dish”

    The problem being, that we’d need a Petri dish for every non-random sample such as occurred in the DP case. :/

  54. Dave_Geologist says:

    Which is why, Everett, you need to “know what’s generalisable and what’s not. And that what’s generalisable for one purpose is not for another.”.

    Of course there will be unknown unknowns. But also known knowns and known unknowns. It’s not perfect, but it’s better than guessing and the more knowledgeable the person making those judgements, the more likely they are to be the right judgements, or at least the closest-to-right judgements.

  55. Ben McMillan says:

    The main interesting thing about the PD data is that it tells you how many were actually infected, because of systematic testing, and you can estimate how many infections you have per hospitalisation/death. Not ‘how fast does it spread’, which doesn’t generalise.

    Sure, there are issues generalising the data, but this is still one of the best pieces of information for certain purposes. All data has serious issues, and you shouldn’t look at any one piece in isolation.

  56. Joshua says:

    Ben –

    > and you can estimate how many infections you have per hospitalisation/death. Not ‘how fast does it spread’, which doesn’t generalise

    Do you think that the hospitalizations/death per infection generalize? I would think that the treatment conditions, SES status of the sample, etc., would suggest not.

  57. Joshua says:

    Steven *

  58. John Hartz says:

    The Trump-Fox News propaganda machine is still going full blast. Americans are needlessly suffering and dying because of the poppycock it spews forth.

    The Trump-Fox News feedback loop now going after Andrew Cuomo and Dr. Anthony Fauci, Opinion by Dean Obeidallah*, CNN, Apr 6, 2020

    *Dean Obeidallah, a former attorney, is the host of SiriusXM radio’s daily program “The Dean Obeidallah Show” and a columnist for The Daily Beast. Follow him @DeanObeidallah.

  59. Joshua says:

    Steven –

    In empirical evidence that wearing (non-certified) makes is counterproductive:

    https://slatestarcodex.com/2020/03/31/ssc-journal-club-macintyre-on-cloth-masks/

  60. Joshua says:

    Arrrggghh. Masks not makes.

  61. Ben McMillan says:

    Joshua: DP data won’t accurately generalise. But that means understand the inaccuracy, not disregard the data, if there isn’t much better around.

  62. anoilman says:

    Keep your distance,
    Wash your hands,
    Don’t touch your face,
    Avoid plague ships.

    Masks don’t work. Unless you’re in a dangerous environment (heath care), and you’re properly trained (nurse, doctor), then there isn’t much point. Folks are supposed to follow standard droplet protocol especially in health care, all the time since you don’t know the state of any incoming patient. Dentists don’t like Hepatitis am I right? (Wash your hands, don’t touch your face.)

    First.. you need to put Masks on and take them off safely (training), and if you don’t know how, then you run a risk of infecting yourself by touching the contaminated surface of your mask. An extreme example is Ebola which they test your ability to put on and take off gowns without help, and without ever touching its surface. Failure has obvious consequences.

    Second, the mask doesn’t block everything (as per Joshua’s link), and in many cases its really just a question of how much needs to get through to spread a sickness. Not an equivalent example, but Norovirus needs fewer than 20 particles to spread itself. (Fecal oral transmission, so pretty gross.) How good is that mask precisely? Are there gaps around the edges? How fine is the mesh?

    Asians love their masks… but I wouldn’t read too much into that. China’s wet markets are serving their Traditional Chinese Medicine habit. Want a mask with that? Personally I’d look at the data rather than the opinions.

    I did hear one Chinese doctor saying that the masks prevent spread when you (an infected person) coughs, and you are the one wearing the mask, however the ability of your mask to catch your sputum is equally debatable. Got any data on that? In the West, the recommended method is to cough into your arm pit and catch what you can, and not spread it. Have I mentioned washing your hands yet? Wash your hands.

  63. Joshua says:

    Oilman –

    I have a different take – given that there is some evidence that COVID-19 might be spread from being aerosolized (you mention only droplet protocol).

    There seems to be some evidence that severity of the infection is to some degree a function of the magnitude of the exposure: Thus, masks that work imperfectly might reduce the magnitude of exposure to aerosolized virus, and thus, reduce severe reactions (less severe exposure – your body fights off the infection; more extensive exposure your body is overwhelmed and you get seriously ill).

    Of course, there is the potential of problems from false confidence arising from the use of masks. There are some who say that wearing masks increases face touching, others who say it reduces it. Who know, that might depend largely on the individual?

    Sure, keeping distance, washing hands, not touching your face are all very important and first order prevention habits – but they aren’t mutually exclusive with erasing a mask when going outside in public spaces is unavoidable.

    At any rate, I think the confidence of your advice is misplaced.

  64. Joshua says:

    Heh. Washing your hands…. Can’t blame that one on autocorrect.

  65. Joshua says:

    Ugh. Using a mask not erasing a mask.

  66. JCH says:

    In Hubei Province, just everybody in a quarantine-period photograph is wearing a mask. The China CDC simply knows more about this disease and fighting it than any other country. In a province with 58.5 million people, which is about the size of Kansas, the stopped this disease in its tracks at ~68,000 cases, and beat it down to zero new cases for what is now weeks.

    And people keep saying South Korea. I don’t get it. Today there are 31 US states and DC with more than 1,000 cases. No provinces outside of Hubei had more 1500, a tiny number had more than 1000, and the vast majority had less than 1000.

    The DC are has more than 5000.

    The west has done a horrible job. I don’t listen to their experts.

  67. izen says:

    @-Joshua
    “There seems to be some evidence that severity of the infection is to some degree a function of the magnitude of the exposure:”

    AFAIK (limited!) the evidence is that the viral load you develop is related to the severity of the infection.
    But it seems unlikely that is directly related to the infective dose. With most viral infections it is not dose related, a few hundred viruses (viri?) are sufficient, just one that co-opts the cellular machinery will then make millions. It is more likely that the efficiency of your immune response at the initial infection level, how well the system recognises and kills your own infected cells before it develops antibodies to the virus is a key factor in determining the viral load.
    That is down to a multitude of incredibly complex individual factors within each individual’s interferon A responses and more in the immune system. It all gets FAR more complex that simple energy based dynamical systems like weather and climate…

    While it is unlikely that masks are totally useless, I suspect the degree that they are protective falls below the ‘noise’ level if you did a carefully controlled blind randomised trial – even if such a thing was possible. There may be a small benefit in preventing a carrier from spreading droplets from coughing, talking and breathing, but an effective testing system to identify positive cases early, while asymptomatic but contagious would be a better bet. That is why the contact tracing and quarantine practised by S.K. Taiwan, Iceland, etc has probably worked.
    Quick, accurate and very widespread repeated testing of most of the population is required with isolation, if you cannot do the location, identification and tracing of those Nations. Without any effective clinical treatment of severe cases it is the only tool available until a vaccine, specific anti-viral or herd immunity can play a role.

  68. anoilman says:

    Joshua… I’m by no means an expert on this material… My wife is. This is all she studies. This is all she analyzes. Everything you’re all saying here, I’ve been hearing for the last 20 years. All the data.. all the viruses.. Chinese Wet Markets… all of it.

    JCH… China’s first response to COVID-19 was to ignore it, and attack anyone who spoke out about it. (I don’t think you’re saying we should do that.. but you should know that.) With a boot for every neck, it was easy for China to shut things down after they totally failed.
    https://www.theguardian.com/world/2020/mar/20/chinese-inquiry-exonerates-coronavirus-whistleblower-doctor-li-wenliang

    In the west its dicey to say our politicians actually listen to experts either. I don’t think that would be a stretch to understand for anyone on this forum.
    https://www.desmogblog.com/2020/03/16/climate-science-deniers-downplayed-covid-19-cato-acsh-aei

    Speaking of expert advice taken, how’s that UK herd immunity thing coming along? Is it going as well as the experts expect? At least the nurses aren’t trying to downplay washing hands like they used to.
    https://www.economist.com/britain/2007/10/18/now-wash-your-hands

  69. Joshua says:

    izen –

    I didnt use the term “viral load” for that reason.

    I get your point but I’m not sure you’re right.

    https://www.newscientist.com/article/2238819-does-a-high-viral-load-or-infectious-dose-make-covid-19-worse/

    Do you have some links?

  70. Joshua says:

    izen –

    Also, I fail to understand why you think that (particularly widespread) use of masks wouldn’t significantly reduce transmission through droplets. They’d clearly be a barrier for that, if not for aerosolized particles.

    Of course, things like testing and contact tracing would have a big impact (such a terrible shame that in US we’ve failed in those aspects so miserably) – but that bears no direct relationship to the impact of mask-wearing.

  71. Joshua says:

    Also –

    > There may be a small benefit in preventing a carrier from spreading droplets from coughing, talking and breathing,

    This seems to (kind of?) conflate droplets with aerozolized transmission.

    This guy happens to be a good friend of my brother – perhaps I’m biased but for sure he knows more than I.

    https://www.rollingstone.com/culture/culture-features/should-you-wear-a-mask-to-fight-coronavirus-a-top-doctor-weighs-in-angry-it-has-come-to-this-976620/

  72. Steven Mosher says:

    “In empirical evidence that wearing (non-certified) makes is counterproductive:

    https://slatestarcodex.com/2020/03/31/ssc-journal-club-macintyre-on-cloth-masks/

    huh. I’d call that one confused study

  73. JCH says:

    China’s first response to COVID-19 was to ignore it, and attack anyone who spoke out about it.

    This is an invention of the same people who brought us yellow cake, weapons of mass destruction, weapons grade powder, and the Gulf of Tonkin.

  74. Joshua says:

    Yeah. I agree the study sucks….but it is some empirical evidence.

  75. Joshua says:

    Oilman –

    If you could get some links from your wife, I’d appreciate it.

  76. Steven Mosher says:

    “Keep your distance,
    Wash your hands,
    Don’t touch your face,
    Avoid plague ships.”

    oilman forgot one of the KEY directives that we get here in Korea. I’m not shocked.
    Public health education is abysmal in the west.

    what did he forget?

    Cover your cough.

    It was quite interesting here as one of the civil defense guys failed to cover his cough
    properly on TV. 3 times. Three times! So of course he was exposed on the nightly news.
    The whole nation got to watch tape of him failing to cover his cough properly.

    But note.

    1. Nobody argued that people should NOT cover their cough merely because
    an expert failed to cover his cough correctly
    2. a mask covers your cough without thinking,

    “Masks don’t work. Unless you’re in a dangerous environment (health care), and you’re properly trained”

    Washing your hands doesn’t work, unless you are properly trained. And obviously since the infection spreads in hospitals patients to the staff, it is clear that even medical professionals fail to wash their hands properly every time. And they fail to keep their hands off their face.
    In short, No one argues that you should not wash your hands because it takes proper training
    to wash your hands.

    It is interesting to see the east/west divide –even amongst health experts– on this.
    Pretty much all the giants infectious diseases here in HK, Korea, Singapore, Taiwan and china
    recommend masks. Precautionary principle rules.

    Meanwhile, in the west, lack of adequate masks for health workers leads some western “experts”
    to make silly arguments to prevent a “run” on masks. Anyway, when we had a shortage of
    masks in Korea, the experts were just honest about it and masks were rationed.
    Health workers came first, everyone else, get in line twice a week. The key for me
    was listening to western experts claim there was scientific evidence that wearing a
    mask increased face touching. Nope. That there was evidence that it would make you
    more prone to infection. Nope. meanwhile in the east, the experts are much more honest.
    honest about the limitations of the data, honest about the pre cautionary principle, honest
    about the shortages, and they actually provide training in how to don and doff. Imagine that!

    The key for me was the utterly stupid argument that you should not use a mask because
    you are not trained to. And at the same time seeing no one make that argument about hand washing. I contrasted that with what we see in the east: constant PSA on
    1. How to wash your hands properly
    2. How to don and doff masks properly
    3. How to clean surfaces
    4. How to social distance
    5. How to cover your cough

    part of me want to say shame on western health experts for shading their advice to protect their tribe. Here’s a clue. A reasonable public understands that health warriors on the front lines
    come first. rationing was accepted in china ( two masks a day max) and was accepted in Korea.
    There’s no need to shade your advice about masks or misrepresent the actual lack of definitive science.

    There is uncertainty in all of this, so it’s important to watch how people USE the uncertainty,
    because we are in a PNS situation and people will use the uncertainty in a way
    that preserves their values and preserves their interests and their tribes interests

  77. Joshua says:

    Steven –

    > It is interesting to see the east/west divide –even amongst health experts– on this.

    Having lived in Korea and Taiwan, and worked very closely with many Asian clients and students got decades, I suspect it has something to do with the individualism/collectivism breakdown.

    Of course, it’s easy to over generalize and reverse engineer to find that breakdown as a bias confirming explanation… but I still think it helps explain the different approach to mask-wearing.

  78. Steven Mosher says:

    Joshua yes it’s some evidence.
    Sadly most people in the USA have no idea how to get N95 masks from Mainland or
    HK. The state of New York refused to buy unless they could speak to my friends factory in English.
    FEMA may help solve some of this by putting boots on the ground and doing factory audits

  79. Steven Mosher says:

    “Having lived in Korea and Taiwan, and worked very closely with many Asian clients and students got decades, I suspect it has something to do with the individualism/collectivism breakdown.”

    I read an interesting piece on it relative to “Christian” values. Veils and truth and all that.

    hmm

    https://www.nationalreview.com/2020/03/coronavirus-medical-masks-western-world-unlike-to-adopt-long-term/

    The question for china is will they change the cultural attitude toward spitting
    which is seen as healthy

  80. JCH says:

    Where it appears everybody was forced to wear a mask, the red are numbers total confirmed cases for the province::

    Versus where people were discouraged from wearing a mask, and these numbers are all going to get a lot bigger:

    New York – 131,916; New Jersey – 41,090; Michigan – 17,221; California – 16,019; Louisiana – 14867; Massachusetts – 18,837; Florida – 13629; Pennsylvania – 13,127; Illinois – 12,262; Washington – 8,326; Texas – 8088, etc.

    Kansas is a little bit bigger than Hubei.

    They say the US attitudes about spitting changed after the 1918 Spanish Flu.

  81. Steven Mosher says:

    Another western expert changes their advice

    https://nationalpost.com/opinion/jonathan-kay-the-case-for-wearing-face-masks-during-covid-19-was-obvious-a-month-ago

    tam was one of the worst examples I found of actual scientists saying questionable shit
    about masks. Makes you wonder whether experts should be able to say shit about policy
    without help from scicom people

  82. anoilman says:

    Mosher… I did mention coughing.. read further down.

    Joshua: My wife is says… go to the WHO, and read. She also says there’s going to be a lot more journal papers on masks coming out soon. (No.. she isn’t getting any for us, and we’re both high risk for getting unplugged.)
    https://www.who.int/emergencies/diseases/novel-coronavirus-2019/advice-for-public/when-and-how-to-use-masks

    JCH: There’s more to that graph than masks. Like boots to necks, when China forcibly took control and stopped COVID. China instantly built hospitals to deal with influx of patients, and after they locked everyone in their homes, they engaged in aggressive transmission tracing. (We still aren’t doing any of that.)

    Koreans? They got spanked by SARS and MERS, so they were prepared. They knew they had to LEAN on the religious gathering for contact tracing. Finally, Koreans had run pandemic drills just last December.

    What were we doing last December? Watching Trump not get impeached? What were we doing in January? Trump said its totally under control in the USA while China was applying boots to necks. (In February Spain was telling people that they shouldn’t go to gatherings, but that they wouldn’t stop them. Italy was just getting its first cases, and hadn’t done much but stop flights.)
    https://www.washingtonpost.com/graphics/2020/politics/trump-coronavirus-statements/

    How about Japan? I mean.. its crowded, and those dang teenagers are hanging at the Cherry Blossoms again. You see about 10% of the people wearing masks there. But dang the place is clean. Even the outdoor markets smell… clean.
    https://www.cbc.ca/news/world/japan-covid-19-strategy-1.5520387

    Japan is still steady cases doubling every 10 days;
    https://www.visualcapitalist.com/infection-trajectory-flattening-the-covid19-curve/

  83. John Hartz says:

    As detailed in the following article, supercomputers now have a key role to play in battle against COVID-19.,,

    The supercomputers will run a myriad of calculations in epidemiology, bioinformatics and molecular modeling, in a bid to drastically cut the time of discovery of new molecules that could lead to a vaccine. Having received proposals from all over the world, we have already reviewed, approved and matched 15 projects to the right supercomputers. More will follow.

    Inside the Global Race to Fight COVID-19 Using the World’s Fastest Supercomputers</strong

    The director of IBM Research explains how the COVID-19 High Performance Computing Consortium came together in just a few days

    by Dario Gil , Scientific American, Apr 6, 2020

    https://blogs.scientificamerican.com/observations/inside-the-global-race-to-fight-covid-19-using-the-worlds-fastest-supercomputers/

  84. Steven Mosher says:

    “Versus where people were discouraged from wearing a mask, and these numbers are all going to get a lot bigger:”

    I left for the USA ( Seattle) on Feb 12 and returned on the 15th.
    Flight out? everyone had a mask. Flight back? I was the only person. not even the aircrew was
    equipped.

    on average about 1% of people on aircraft are infected. This is a bit hard to figure as you have to
    hunt for the denominator. Countries like Korea report very selective stats that
    serve a political agenda.

    they report Imported cases and break out NATIONALS versus Foreigners , but not
    total number of travellers. It’s pretty obvious why. In China they reported nationals versus
    foreigners, and while the cases of Foreigners is always a tiny number (relative to nationals)
    they are now banning all foreigners. Taiwan did this early.

    The airport arrival data ( and selective reporting) is interesting
    In Korea 8% are Foreigners 92% nationals.
    Flight departures from USA and EU are 90% of the infections
    40% are caught at the airport, which means they presented with fever ( how the fuck
    is the US and EU not pre screening)
    60% are caught in quarantine. which is actually enforced with fines and jail terms if you
    violate ( they just deported a stupid Brit)

    To see how competent public health officials work
    see what Taiwan does, day by day for the crisis

    https://cdn.jamanetwork.com/ama/content_public/journal/jama/0/jvp200035supp1_prod.pdf?Expires=2147483647&Signature=bIZCLS7ZLWTJd~U~H40JgiEGdFb3ggVUJpBvJ7KdANK7HgK1zaj4uWHvqweGym1nWfO~nXt9Y5i1vX79pF7zjjqfzmJAy3udTdpVVZQe07xnQIPcBMXLwZ5XjgTO8yKFXVIpxsXhrmOu8sGSpKiEmQ86ZCKfOTar7fMAGmUCtjiYVFwf31K3REWAA-r3hZyoZpqz3QKpVgpsRpF9fV9thQCq0~yvbvRKTH4PcoB~CZgmXH7rpVb6bILXQn5zBCphf6pyLAa4zIebUEKfCdCYdSdi9LeIEUsesqsYpNWgHJcr4K1LC0hFlst0RHQz-vZ7I-OvrX~5jel6zjjtuDQzjQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA

  85. anoilman says:

    Mosher: There is incredible pressure to say and do things right now on health care professionals. Unbelievable pressure. In many cases you have politicos and nontechnical executives pulling strings and ‘declaring it so’.

    Hmm… In your link, Dr Tam says the masks will not protect you from getting sick.

  86. Joshua says:

    > Sweeping mask recommendations—as many have proposed—will not reduce SARS-CoV-2 transmission, as evidenced by the widespread practice of wearing such masks in Hubei province, China, before and during its mass COVID-19 transmission experience earlier this year. Our review of relevant studies indicates that cloth masks will be ineffective at preventing SARS-CoV-2 transmission, whether worn as source control or as PPE.

    OK. So maybe wearing masksndont make a difference in countries where people spit a lot, and live in massive apartment complexes, and before there was social distancing. And from the we can determine that….??

    However, lots o’ references to empirical studies.

    https://www.cidrap.umn.edu/news-perspective/2020/04/commentary-masks-all-covid-19-not-based-sound-data?

  87. Steven Mosher says:

    ‘Japan is still steady cases doubling every 10 days;
    https://www.visualcapitalist.com/infection-trajectory-flattening-the-covid19-curve/

    japan has a policy of dont test too much, they do testing at a pathetic 365 per million.
    One reason is that positive cases must be hospitalized by law.

    Take Korea as an example:
    1. if you test positive you go home or too one of the 7000 comandeered buildings
    that serve as “life health centers”
    2. if you get bad, you go to the hospital
    Consequently, Korea can afford to contact trace and test, as not all positives go to the hospital
    In japan, by law, all positives must be hospitalized.

    And dont forget, Japanese and Koreans dont shake hands. Boris should have bowed

  88. Steven Mosher says:

    Joshua
    ‘Martin and Cowling come down on opposite ends of the face mask debate. Martin, who said she has watched misuse of masks and gloves spread disease in emergency situations, believes masks can create a false sense of security.”

    read that and ask yourself, what kind of expert would give that as evidence?

    ‘Martin and Cowling come down on opposite ends of the had washing debate. Martin, who said she has watched the failure of proper hand washing spread disease in emergency situations, believes hand washing can create a false sense of security.”

    I mean seriously. And expert who uses that kind of logic.. Hey! Condoms break! therefore
    they give a false sense of security!

  89. Joshua says:

    Steven –

    Click through to the “anti’ articles.

  90. izen says:

    @-Joshua
    “Also, I fail to understand why you think that (particularly widespread) use of masks wouldn’t significantly reduce transmission through droplets.”

    I am well out of my lane here, just another interested observer with a long-term curiosity about genetics and how that is expressed in disease and immune systems.
    I would concur with Mosher that the contradictory and ambiguous advice on masks, and many other aspects of this pandemic have more to do with local social tradition, experience and political expediency than any firm scientific data from which we can draw conclusions.

    But AFAIK there is little evidence that the severity of a viral infection has much correlation with the size of the infective dose.. Although it may have an impact on the likelihood of catching the disease. Most research into the minimum infective dose has problems, but tends to indicate that at most only a thousand or so viri are needed to cause and infection. But the individual immune response dominates the subsequent severity.

    https://www.researchgate.net/publication/227225392_Minimum_Infective_Dose_of_the_Major_Human_Respiratory_and_Enteric_Viruses_Transmitted_Through_Food_and_the_Environment
    “Notwithstanding these limitations, the MID of respiratory and enteric viruses appears to be low and should be viewed in relation to the likely host characteristics of the at-risk population of interest.”

    There is very little good evidence for the route of transmission of COVID19, but some indication that surface contamination is at least, and possibly more, significant than aerosol inhalation. In that case it is hand to mouth/nose contamination that is important rather than breathing. So if you in the presence of a coughing case then a mask may be of some protection. But in general WASH YOUR HANDS is probably a better precaution.

    I have a suspicion that this focus on masks, along with the enthusiasm for chloroquine and burning down 5G masts has more to do with people wanting to have the (probably fictive) feeling they have some sort of control, a means of individual autonomy, over their fate in this situation. Faced with the prospect that catching the virus is random but likely, and the subsequent outcome is also random, but skewed towards the old and infirm with no effective medical intervention available I think people grasp for some means of commanding their own destiny.
    However illusory.

  91. Joshua says:

    > But AFAIK there is little evidence that the severity of a viral infection has much correlation with the size of the infective dose..

    Did you read the article I linked above?

    > For influenza, a higher amount of virus at infection has been associated with worse symptoms. It has been tested by exposing volunteers to escalating doses of influenza virus in a controlled setting and carefully monitoring them over several weeks. This hasn’t been done with covid-19, and is unlikely to happen, given its severity.

  92. izen says:

    @-Joshua
    “Did you read the article I linked above?”

    Yes, I had already seen the NewScientist article and the research paper it links to –

    https://www.ncbi.nlm.nih.gov/pubmed/25416753

    Note that this found a small difference in outcomes, with overlapping error bars for an order of magnitude difference in infective dose.

    I am not discounting the possibility of a proportional relationship between infective dose and disease severity.
    But I would expect from the pattern of behaviour with most viral infections that the infective dose has a much bigger role in the probability of actually developing the disease. But the severity of the disease is dominated by the host response.

    Personally I am happy to wear masks if available and even happier to wear gloves which I can change and discard frequently.
    But I do not have any strong expectation this will do anything significant to alter the course of any infection I catch, just reduce the chances of catching it. But even to achieve that end both masks and gloves need to be safely removed and replaced at least after every encounter/ excursion in an ‘at risk’ situation.

  93. dikranmarsupial says:

    “they give a false sense of security!”

    ISTR Jeremy Clarkson once suggested getting rid of seat belts and putting a sharp metal spike on the steering wheel in order to make people drive more carefully. I’d be skeptical of an argument analogous to one used by Jeremy Clarkson. ;o)

    Personally, where the scientific evidence is not unequivocal, and where my intuition is not great (i.e most things), I’d align myself with relevant scientific bodies, such as the WHO, rather than focus on individual studies (which is likely to reflect existing cognitive biases). There is no scientific position that is so nutty that you can’t find papers to support it. For example, should we believe Salby or Berry, or Essenhigh about the cause of the rise in CO2, or should we believe the IPCC?

    wonders if duct-taping some kitchen towels on your face would work… ;o)

  94. Steven Mosher says:

  95. anoilman says:

    I think one should also look at what the health care system and other government systems are that are in place. In China… health care is a mess (if you’re confused, then I’d prescribe more Rhino horn), but they have an army and plenty of boots to place on necks.

    In Korea its public\private, and UK… public private, etc. Canada, is Public only. For instance, in Canada, each province is a separate health care system broken into regions, and mostly all the hospitals in a region fall under a set of practices. (Alberta is the exception.. its a provincial service, and provincially controlled.)

    US healthcare is a Munted Sh*t Show, which has always made me wonder how it could handle a situation like this. Each and every hospital may fall under its own control and jurisdiction or authority. Even within a city, each hospital will have its own arrangement with different health care providers, so you can’t go to the nearest one, you have to go to the one you paid for, and receive the service you’re allowed to receive. You have different state and federal regulations poking into that as well. With so many money wranglers, and political interference, how does that pull together into a cohesive system to deal with a crisis like a pandemic?

  96. Steven Mosher says:

    Cowling was just on the news here, here’s his paper

    https://www.nature.com/articles/s41591-020-0843-2

  97. John Hartz says:

    A must-read analysis — especially for all of the armchair COVID-19 experts posting comments on this thread…

    We’ve grown accustomed to living through an information war fought largely by hardened political operatives and trolls. But while the coronavirus crisis is political and will continue to be politicized, its most consequential fights will take place in the “fog of pandemic” where so much of our data — from health statistics to economic indicators — is flawed or evolving. Today’s propaganda could be tomorrow’s truth. Or vice versa. Even the good guys are working with limited information and hoping for the best. We are not prepared for what’s coming.

    The best illustration of this challenge is the changing consensus and public messaging on wearing masks.

    What We Pretend to Know About the Coronavirus Could Kill Us, Opinion by Charlie Warzel, Sunday Review, Apr 3, 2020

    PS – As you might suspect, much of Warzel ‘s analysis is applicable to the never-ending climate propaganda war conducted in the cybersphere.

  98. John Hartz says:

    More reality….

    Coroners worry Covid-19 test shortages could lead to uncounted deaths by Blake Ellis, Melanie Hicken and Ashley Fantz, CNN Investigates, April 7, 2020

    https://www.cnn.com/2020/04/06/health/coronavirus-coroners-uncounted-deaths-invs/index.html

    Based on the news reports I have read/seen, many countries are (intentionally or unintentionally) under-reporting the number of deaths due to complications caused by COVID-19. infections.

  99. “Wearing a medical mask is one of the prevention measures that can limit the spread of certain respiratory viral diseases, including COVID-19.”

    How at risk are you to SARS-COVID-19?
    Lymphopenia predicts disease severity of COVID-19: a descriptive and predictive study.
    Of course, a small study not yet reviewed.
    But % lymphocytes roughly corresponds with disease severity and outcome.
    If your lymphocyte percentage, available from most CBCs, is less than 25%, you may be at greater risk.

    This is roughly consistent with immunosenescence.
    Evidently, immunosenescence increases with age occurs in most species.
    One cannot reduce one’s age.
    However, immunosenescence also increases from the results of the modern processed diet, which we are not evolved for.
    Reversing metabolic syndrome may improve immune response, including the potential response to vaccination.
    Given the global epidemic of metabolic syndrome, eating a nutrient dense, whole foods diet may help improve one’s health span including response to COVID-19.

  100. David B Benson says:

    https://medicalxpress.com/news/2020-04-covid-average-actual-infections-worldwide.html

    Highly underreported. Maybe 20 times the reported number.

  101. izen says:

    @-TE
    “…eating a nutrient dense, whole foods diet may help improve one’s health span including response to COVID-19.”

    But ‘junk’ foods like soda and ready meals packed with sugar and fats provide more calories per unit cost for the consumer, at a greater profit margin for the producer.
    A win-win for both sides. (/s)

  102. An older code says:

    @Dikran

    I think the steering wheel quote was from Peter Wheeler, ex owner of TVR

    A contrarian by nature and deeply suspicious of car safety systems

    “At the same time, he made TVR infamous by turning his back on such de-rigueur safety equipment for his light, yet immensely strong, cars as air-bags and anti-lock brakes. The former he thought unsafe at any speed, the latter unnecessary in a properly set-up car. Famously, he said that TVRs were built to be “relatively safe upside-down, proven by customers”.

    https://www.theguardian.com/business/2009/jul/02/peter-wheeler-obituary

  103. anoilman says:

    TE: One of the best ways to survive a pandemic is to not get sick in the first place. (Have I mentioned washing your hands yet?)

    Pop quiz! What is the compliance rate for medical professionals to wash their hands? Anyone? Anyone? Beuller? Yup! 70% ish… Cool huh!

    Click to access who_guidelines-handhygiene_summary.pdf

    Izen: I like salt.

  104. Steven Mosher says:

    “Highly underreported. Maybe 20 times the reported number.”

    read the “paper”
    its junk.

  105. Steven Mosher says:

    Hilarious stuff from the WHO

    ‘Medical masks should be reserved for health care workers.
    The use of medical masks in the community may create a
    false sense of security, with neglect of other essential
    measures, such as hand hygiene practices and physical
    distancing, and may lead to touching the face under the masks
    and under the eyes, result in unnecessary costs, and take ”

    ZERO citations for this.
    1, there is no observational evidence that wearing a mask creates a false sense of security
    2. There is no evidence it leads to less hand washing
    3. there is no evidence it leads to touching your face under the mask
    ZERO
    Now, if that were a science paper you would expect a citation. where is the data on this?
    there is no data. Zero data. No observations and no attempt to gather what actual data there is.

    What evidence there is is this. People touch their face about 20 time an hour
    60% of the time in their mouth nose or eyes.
    The real issue is supply. Prepared nations have no issue and of course rationing may
    have to be used.
    A surgical mask costs about .03 to produce. They are available, millions per day easily ramped.
    N95? a little bit more ~1 buck, again, millions per day. But you have to plan. Opps.

    As for the observational evidence on face touching. It’s available ( face touching under masks)
    but it will require folks looking at CCTV footage ( with AI of course) of places where some of
    the population wears masks

    what would that look like?
    For example: in this video you see a mixed population, mask and no masks

    in the beginning you will see a girl tugging at the bottom of her mask.
    and seconds later you will see a guy without a mask rub his nose.
    Further you will see that the folks who don’t wear a mask touch the public surfaces
    ( watch 3 non mask wearers touch the escalator hand rail) more often than the masked people.
    This is something I noticed the other day at the mall. Damn foreigners with no masks
    touching the hand rail, mask wearers? Nope, not a single one. Anyway, It would be cool
    to review footage and actually see. do people wearing masks actually reach underneath them
    to stick their fingers in their mouths or touch their face more?
    so rather than conjecture about behavior of mask wearers, maybe folks should look at data.
    it would be hard, but it beats conjecture

    So, the “data’ actually exists but it would require someone to actually look and see
    My little venture out yesterday told me this. About 1 hour of people watching (from a distance) at the market, taxi stand, and train station.
    Zero incidents of people reaching under their masks. Zero. not 1, Zero. zero point zero.
    Absent any published data on people reaching under masks, I’m gonna suggest that ANY observation is better than mere conjecture. My observation? wearing a mask makes people more conscious of other behaviors they need to follow. But hey it’s not my job
    to study this, but I would expect anyone who published guidance would actually TRY
    and collect data. I would expect them to spend some time watching the public.
    FFS, Jane Goodall did not conjecture about damn animal behavior, she fucking watched.

    The issue is making a Health recommendation based on your concerns over supply constraints and spicing it up with unproven concerns about a “sense of false security”
    It’s more honest to say.
    We have no science on these issues, we don’t even want to study it because we are worried
    the public will hog all the masks .

  106. JCH says:

    On March 12 issued a ban for travel to the United States by foreign national who had been in a certain region of Europe within the last two weeks. US citizens were exempt, but were to return through 13 designated airports where thy were to be screened:

    1 – JFK International Airport – 138,863
    2 – Chicago O’Hare Int -13,549
    3 – San Francisco Int – see below
    4 – Los Angeles Int – 17,165
    5 – Seattle-Tacoma Int – 8,384
    6 – Daniel K. Inouye Int – 410
    7 – Hartsfield-Jackson Atlanta Int – 8,818
    8 – Washington-Dulles Int- 8,915 (VA, MD, DC)
    9 – Newark Liberty Int – 44,416
    10 – Detroit Metropolitan – 18,970
    11 – Boston Logan Int – 15,202
    12 – Dallas/Fort Worth Int – 8,262
    13 – Miami Int – 14,747
    ——————————
    297,701 confirmed cases in the cities/areas served by the above airports as of today
    —————————————————————————————————————

    75% of ~394,500 total confirmed cases in the US when I wrote down the numbers.

    Presumably they washed their hands a lot. Airports have lots of sinks. Wonder what the situation would be if each had been given a mask.

    The above is perhaps the dumbest public health move, well, since ever. On the day it was completed their were 2,770 confirmed cases in the entirety of the USA. If people didn’t have COVID-19 when they crowded into airports in Europe, and then boarded crowded airplanes for the long flight home, and then several hours closely packed crowds in the 13 airports, they probably walked out of those terminals with virus in their noses and lungs and on their clothing, and then on to hugging family members who were there to greet them. Others walked down the terminal halls and boarded airplanes to other US airports, and so on. 5 days to incubate, KaBOOM.

  107. anoilman says:

    Mosher.. Personally I think you’d touch your face less with a mask on, but I’m not sure how that would prevent the spread of viruses.

    And yes, there’s tons of science on these issues. Did you read the WHO article referencing all the articles about hand washing? There was a lot right? The way you quality control your personnel is to follow them around and observe. (Its got nothing to do with Jane Goodall, and I’m pretty sure she didn’t invent the idea.) Yes, they do observations before and after training, and they choose study sizes to get statistically significant measurements. Did you see the ‘before’ numbers? In many cases they are quite abysmal huh?

    There’s tons of science on this. The experts are 100% glaring at, and looking through, and working on this, and have been for oh, 100 years or more.
    https://www.webmd.com/lung/news/20030429/can-mask-protect-you-from-sars#1

    Do you think those mandatory masks worked in 1918? At least they could get into big crowds shoulder to shoulder, that would nice to do again. “No gatherings will be considered fashionable unless the attendees are attired in masks.”
    https://www.cnn.com/2020/04/03/americas/flu-america-1918-masks-intl-hnk/index.html

    So far it looks like we’re doing better this time.

    Personally, I wonder about packaging the products I buy right now. Other people touch them;
    https://www.journalofhospitalinfection.com/article/S0195-6701(20)30046-3/fulltext

  108. dikranmarsupial says:

    “Hilarious stuff from the WHO”

    hubris

  109. Dave_Geologist says:

    Who to believe? A self-published paper by a professor of development economics, or dozens or journal preprints and peer reviewed papers from expert epidemiologists who’ve studied cases in circumstances where the asymptomatic can’t hide, at least not in those sort of numbers.

    Tough choice.

    Not.

  110. dikranmarsupial says:

    @An older code – thanks for that, especially the ” “relatively safe upside-down, proven by customers”.

    IVSTR something about TVRs being reverse kit cars, where they are delivered assembled and turn into a kit over tme?

  111. Dave_Geologist says:

    Big macho car ≠ safe car.

    Decades ago, on fieldwork, I passed what I later learned was a fatal head-on collision between a Series 3 Landrover and a family saloon. The saloon occupants were the ones who survived. The ‘drover’s ladder chassis penetrated the front of the saloon, but the crumple zone crumpled and the engine over-rode the girders. The ‘drovers bodywork crumpled and detached from the chassis, and its engine and gearbox were pushed into the passenger compartment.

    My first vehicle was of similar construction, a Morris Minor 1000 ex-post-office van, the heavy-duty version with 5-leaf instead of 3-leaf springs and 1100cc engine (handy for transporting rock samples and camping gear, as was the 6-inch ground clearance). I helped right one that had rolled onto its roof (not mine). The driver survived and walked away, probably helped by the high roof behind. And I presume he was wearing a seat-belt, which was not compulsory in those days.

    The entire rear bodywork was bolted onto the chassis so we helped take it off, turn it over and hammer out the dents in the the roof from the inside. After that it was right as rain. Not even a broken windscreen. Bet he didn’t tell his insurer though!

  112. dikranmarsupial says:

    MiniRant: there should be more stringent driving tests for anybody driving a car of over, say 1.5 tons*. It may increase your survivability in an accident, but it decreases that of anybody you may run into who is driving a more sensible vehicle. With rights come responsibilities.

    * or anything with a 0-60 of less than, say 7 seconds.

  113. JCH says:

    The asymptomatic is an assumption from the seasonal flu mindset. Early results indicate they’re not there. Which, if you grew up treating viral diseases in pastures, and hog barns, and chicken houses, makes sense.

    Airports Reel as New Coronavirus Screening Goes into Effect

    76% of all USA cases, now 400 thousand, are in 12 of the 13 cities/regions served by those airports. Only exceptions is Hawaii.

    If you look at cities with airports not included, they have far fewer confirmed cases: San Diego, Denver, Pittsburgh, Minneapolis St Paul, Kansas City, St Louis, Charlotte, Tampa, Pittsburgh, Oklahoma City, and Phoenix. The only real exception is New Orleans, which had the same scene: Mardi Gras. Pigs in barn, all sick really fast.

  114. Willard says:

    All the lanes belong to Matt King Coal:

  115. Dave_Geologist says:

    Zoe Harcombe: “I am not aware of any debate about the value of masks among healthcare workers – only among the general public. Yet why would masks protect them and not us?”

    Because healthcare workers don’t have the privilege of staying 2m away from their patients and not spending more than 10-15 minutes in a room with their patients? Or indeed staying 2m away from each other and not spending more than 10-15 minutes in a room with each other?

  116. Joshua says:

    Dave –

    > Because healthcare workers don’t have the privilege of staying 2m away from their patients and not spending more than 10-15 minutes in a room with their patients? Or indeed staying 2m away from each other and not spending more than 10-15 minutes in a room with each other?

    Walk me through the mechanics of your thinking for how distancing or time of exposure would mean that masks are effective for healthcare workers but not the general public.

    Seems like they’d like be effective against large droplets, which may not likely (often) travel more than 6 feet. Thus, theyre effective against transmission from large droplets within 6 feet, but have no impact on the pathway of transmission where someone sneezes and a droplet lands in a surface which is later touched by someone else. Is that your reasoning?

    And the mechanics of time of exposure as a predictor for the value of masks (differentially between healthcare workers and the general public) would work how?

  117. Joshua says:

    JCH –

    > The asymptomatic is an assumption from the seasonal flu mindset. Early results indicate they’re not there.

    Could you rephrase that?

  118. dikranmarsupial says:

    “Because healthcare workers don’t have the privilege of staying 2m away from their patients and not spending more than 10-15 minutes in a room with their patients?”

    and the incubation period is quite long so you can be spreading the virus without having symptoms to tell you to self-isolation.

  119. dikranmarsupial says:

    “Walk me through the mechanics of your thinking for how distancing or time of exposure would mean that masks are effective for healthcare workers but not the general public.”

    Say you don’t have COVID-19 but one of your fellow health-care professionals does, but doesn’t know it. They spread it to the patients because they are currently asymptomatic and they can’t avoid being in close contact with the patients. Neither can you, so if the mark doesn’t protect *you* directly from being infected because of the need for close contact, masks being worn by all healthcare profiessions, including your unknowningly infected colleague does.

    Doesn’t apply to general public because we have the option of staying away from people, which does protect us individually.

  120. dikranmarsupial says:

    Hospitals are hubs in the transmission network, doctors and nurses are ideal vectors – they see many patients a day. That would be my intution.

  121. Joshua says:

    > Say you don’t have COVID-19 but one of your fellow health-care professionals does, but doesn’t know it. They spread it to the patients because they are currently asymptomatic and they can’t avoid being in close contact with the patients. Neither can you, so if the mark doesn’t protect *you* directly from being infected because of the need for close contact, masks being worn by all healthcare profiessions, including your unknowningly infected colleague does.

    I’m still not quite getting how the six feet distancing or the time of exposure explain the “mechanics” of how transmission would differ so as to determine the effectiveness of masks.

    > Doesn’t apply to general public because we have the option of staying away from people, which does protect us individually.

    But the ability to stay away from people is not uniformly or universally applicable. Some people need to go out into the public for a variety of reasons. If they have to do so, invariably there will be occasions where accidents will happen, or inattention or sloppiness will happen – and then we have to expand that exponentially. Regardless, that might mean that the need for healthcare workers is more crucial – I think everyone accepts that – but not that there is some kind of line of distinction between the utility of them being worn by healthcare workers from the utility of them being worn by the general public.

  122. JCH says:

    I’m just laughing. You cannot stay 6 feet away all the time, and droplets can travel more than 6 feet. See choir.

    Wuhan:

    ~57,000 cases, ~2100

    New York:

    ~142,000 cases, ~5,100 deaths

  123. dikranmarsupial says:

    “But the ability to stay away from people is not uniformly or universally applicable.”

    There are always exceptions to every rule, and e.g. the WHO cannot be expected to enumerate all of them in guidance to the general public (indeed I suspect that would be deeply counterproductive)

    There are risks associated with using masks as well as benefits, especially reusable home-made ones. For example you put on your mask and go out and then forget to wash your hands before taking your mask of. The next time you put the mask on, you have just indirectly touched your face because your fingers have touched something you then put over your mouth.

    If you don’t currently have COVID-19, and the 2m distancing works, then in that case you have just increased your risk, not reduced it.

    As you say, sloppiness will happen, and it can can increase the risks of using masks as well.

    Now organisations like the WHO will be looking at the big picture and balancing the known advantages and disadvantages in the guidance they give.

    As I pointed out, as I understand it, the main advantage of a mask is that it reduces the risk of you infecting someone else, rather than the risk of them infecting you.,

  124. Maybe we can clarify something. Currently the UK is essentially in lockdown. We’re meant to only go out for essentials, to exercise once a day, and are meant to stay more than 2m away from people who aren’t in our households. Given this, I can see that it may be more important for there to be masks for healthcare workers than for those of us who are not meant to get close enough to anyone else to need one. It could be different if the lockdown were to relax while there was still a good chance of contracting the virus.

  125. dikranmarsupial says:

    “It could be different if the lockdown were to relax while there was still a good chance of contracting the virus.”

    Indeed, I’ve had no difficulty staying 2m from everybody for the last week or so. It might still be the case that it would only help in the sense of reducing the risk of the wearer infecting someone else, rather than reducing their own risk directly.

    The workers that I would place next on the list would be people in the food supply chain.

  126. Joshua says:

    Anders –

    > Currently the UK is essentially in lockdown. We’re meant to only go out for essentials, to exercise once a day, and are meant to stay more than 2m away from people who aren’t in our households.

    I don’t know what it’s like in the UK, but in the US if you go shopping in a supermarket, for essentials, it is next to impossible to uniformly maintain a six foot distance from everyone in the store – even if you fully intend to do so. And of course, there are all manner of essential workers who might be able to reduce probabilities by staying six feet away from everyone, but will nonetheless not be able to do that all time if only because they can’t control the behavior of others.

    > Given this, I can see that it may be more important for there to be masks for healthcare workers

    Yes, of course. And unfortunately, the reality on the ground is that at some level we have to make choice in that aspect because of the shortage of masks.

    But, that doesn’t speak to the question of the effectiveness of home-made masks, and whether there should be an all out effort to get people to use them.

    And part of the problem with this discussion is that we’re not really distinguishing the types of masks we’re talking about.

  127. Joshua says:

    But none of that really answers my question, which is about an explanation of the “mechanics” for why a mask would be effective within six feet or with more than 10 minutes of exposure, but not effective beyond six feet or with less than 10 minutes of exposure.

  128. Joshua says:

    > There are always exceptions to every rule, and e.g. the WHO cannot be expected to enumerate all of them in guidance to the general public (indeed I suspect that would be deeply counterproductive)

    I’m not really referring to exceptions to the rule. My experience, and that of many people I’ve talked to, is that it is impossible to stay six feet from everyone when performing essential tasks in public. For that reason, as someone considered at high risk, I have eliminated going out into public. But I can do so because of being relatively privileged in some respects. I’m not questioning the value in trying to apply the six foot rule, or that in doing so you can reduce the relatively level of risk. I’m asking how the application of the six foot rule would somehow mean a binary assessment of the value of the effectiveness of masks.

    > There are risks associated with using masks as well as benefits, especially reusable home-made ones. For example you put on your mask and go out and then forget to wash your hands before taking your mask of. The next time you put the mask on, you have just indirectly touched your face because your fingers have touched something you then put over your mouth.

    Yes, I realize that. But not of that speaks to the questions I’m asking.

    > If you don’t currently have COVID-19, and the 2m distancing works, then in that case you have just increased your risk, not reduced it.

    Of course distancing reduces probability. I would avoid such binary statements as distancing works, except as an overall statement of the overall trend.

    > As you say, sloppiness will happen, and it can can increase the risks of using masks as well.

    Of course. So then the question is an assessment as to whether the increased risk from sloppy use of masks outweighs any potential benefit that wearing masks, while not eliminating risk, might reduce probabilities – a reduction which would then get magnified exponentially.

    > Now organisations like the WHO will be looking at the big picture and balancing the known advantages and disadvantages in the guidance they give.

    I don’t question that. But this is all, also, a moving target. The CDC is reevaluating as we apeak.
    Additionally, it’s hard do know how much of their assessment is based on the need to provide access for critical care workers. That may or may not be relevant to the use of home-made masks.

    > As I pointed out, as I understand it, the main advantage of a mask is that it reduces the risk of you infecting someone else, rather than the risk of them infecting you.,

    Again, I already know this. So that is exactly a reason why the public wearing a mask when out in public, irrespective of whether they’re able to maintain a 6 foot distance all the time, would be recommended, as long as them doing so wouldn’t reduce access for healthcare workers.

  129. Joshua says:

    And with that, dirkan, I’ll leave off further interacting with you on this.

    My original question was directed towards Dave. I’ll leave it to him to answer if he sees fit.

  130. dikranmarsupial says:

    “But none of that really answers my question, which is about an explanation of the “mechanics” for why a mask would be effective within six feet or with more than 10 minutes of exposure, but not effective beyond six feet or with less than 10 minutes of exposure.”

    the point being made is that it is not the mechanics by which it protects the wearers, but the mechanics by which it protects others from the wearer. If you are >2m away, then that protects others from your exhalations, if they are closer, then they need more protection from you, for instance putting a mask as close to the output port as possible,

    If we have a radioactve source, then I can let a few cm of air protect me from alpha/beta radiation. If I need to be closer than that, perhaps put some lead shielding in between.

  131. dikranmarsupial says:

    “And with that, dirkan, I’ll leave off further interacting with you on this.

    now that is just rude.

  132. Joshua says:

    > now that is just rude.

    I apologize. No intention to be rude. But given what went down thus far, I suspect further interaction won’t be of value to me. You can put that on me.

  133. izen says:

    One of the more depressing aspects of this pandemic apart from the death toll skewed towards the older and infirm, is the inherent human desperation to reframe any event into a narrative of heroes and villains

    Contagious disease has its own dynamic based in the path and probability of cross infection. The ways of breaking those chains of communication are quite well known from quarantine since time immemorial, to washing your hands since Semmelweis

    In the current pandemic the heroes are chlorquinine, mask wearing and ICU/respirator provision. None have significant evidence they are of much benefit.
    The villains are China, who may have wished for the best in the early stages, but could not have definitively declared it a highly contagious pandemic risk before the full pattern of spread and morbidity was established.
    Then there is the totally delusional conspiracy theory that there is no virus, it is 5G radiation making people sick, or at least weakening their purity of essence to make them susceptible to a bug.

    People who break the rules on social distancing are also attacked and governments congratulate themselves on stopping air travel. But most stopped travel of nationals from selected countries, but encouraged the return of their own citizens and subjects from those same ‘at risk’ regions because to tell a US citizen or UK subject they HAD to stay in S Korea or on a cruise ship was politically untenable.
    And while a large majority of the population can self-isolate and practise social distancing, there is a basic logistical problem that a certain proportion of the working population and their children HAVE to travel and mix as key workers just to keep the health care, utilities, and food productions/supply system operational for the rest.

    The Nature of infectious disease really does not lends itself to any characterisation of villainy, it lacks any moral intentionality that would make the concept meaningful.
    The only heroes in this situation, as with measles, polio, smallpox and flu are thepeople and medical programs that develop vaccines and make them widely and freely available as rapidly as possible.

  134. dikranmarsupial says:

    Let p_i be the probability of using a mask in a risky way resulting in exposure. Let p_a be the probability of wearing a mask preventing infection and p_e be the probability of airborne exposure. If p_i > p_a*p_e then on average, wearing a mask is counter-productive.

    I don’t know what p_i, p_a and p_e are. I suspect the WHO have a much better idea what they are than I do, so I take their advice and assume that if it seems counter-intuitive to me that they probably know something I don’t.

  135. Joshua says:

    I will again post two articles where a close friend of my brother’s is interviewed. I may be biased in his favor, but I find what he has to say to be relevant and I tend to trust his expertise due to knowing his background. This is in addition to the tons o’ material I’ve been reading on both sides of this issue. It isn’t as if he’s some kind of an outlier.

    Some like Zoe Harcombe, I tend to be more skeptical of because of her heavily cultivated public profile.

    https://www.rollingstone.com/culture/culture-features/should-you-wear-a-mask-to-fight-coronavirus-a-top-doctor-weighs-in-angry-it-has-come-to-this-976620/

    https://www.theatlantic.com/health/archive/2020/04/coronavirus-pandemic-airborne-go-outside-masks/609235/

  136. dikranmarsupial says:

    “I apologize. No intention to be rude. But given what went down thus far, I suspect further interaction won’t be of value to me. You can put that on me.”

    that is a classic non-apology apology which merely doubles down by explicitly re-issuing the insult (saying that my contributions are of no value to you).

  137. dikranmarsupial says:

    AFAICS nothing bad had “gone down thus far”, just me trying to answer your question and us not achieving agreement, yet.

  138. Joshua says:

    izen –

    > None have significant evidence they are of much benefit.

    From what I’ve seen that well be true. I’ve been looking a bit at the evidence (I found the discussion of the meta-analyses in the Harcombe piece to be useful, and I found the discussion in the 538 piece to be useful because it referenced meta-analyses) .

    On the other hand, from what I’ve seen the evidence that they aren’t of much benefit seems to fall pretty far short of “significance” as well.

    So what to do in such a situation?

  139. Joshua says:

    dirkan –

    Last comment. I am not getting any value from this discussion with you, and so I wish to discontinue the interaction for that reason.

    You may consider my saying that as an insult. But I don’t see saying that as an insult and it certainly wasn’t intended as an insult.

  140. dikranmarsupial says:

    Repeating the offending statement again – classy.

  141. anoilman says:

    JCH: That’s not how they got things under control. They deployed the military and locked the place down. So here’s a theoretical question for you. If they welded your door shut locking you in your apartment, would you need a mask for being in public?

    If I told you that you’d be hauled away in the streets like dog if you didn’t wear a mask, would you do it? Honestly, I’d bark like a dog to that choir to avoid the humiliation, wouldn’t you? WOOF WOOF! Arooooo!!!!! Gawd I need a mask to avoid that don’t I?

    30 minutes in you can see authorities hauling people away. 31:10 welding doors shut so people can’t get out.

  142. dikranmarsupial says:

    If you don;t think that telling someone who is trying to answer your question that you are not getting any value from them is not an insult, then you have an odd idea of civil discussion. Especially if you do so immediately after a response that didn’t actually address what was said to you.

  143. anoilman says:

    Guys: I’m not sure why we’re all getting excited about this. Yes it sucks.. But no, there’s no data showing that you need masks in public. If you want to wear one, go ahead. Otherwise, there’s no need to get all passionate about it. (Me included. I’m being a bit of a dip.)

    I do find the similarities in argumentation to global warming interesting. It seems that some of us are flipped in perspective and trying to be the experts without anything to back it up.

    From my perspective, in understanding global warming, its all mature science that’s well established and reviewed. It would take an awful lot to knock that out of place. At best, any future work in the field can only augment our understanding.

    Given what I just said, then yes.. maybe masks do help and when there’s studies showing that, it become incorporated into our understanding;
    Keep Your Distance,
    Wash Your Hands,
    Don’t Touch your face, and,
    Wear a mask?

  144. dikranmarsupial says:

    ” If you want to wear one, go ahead.”

    absolutely, but if it is a reusable one, be very careful to avoid contaminating it when taking it off and putting it on again (and storage).

    I’m not particularly passionate about whether to wear masks or not, just on following advice from people who know what they are talking about when I don’t. If the WHO or NHS say “start wearing masks” I will start wearing a mask.

    N.B. AFAICS the WHO are not saying “don’t wear a mask”, just that most of us don’t need to, so if you want to wear one, nobody is saying you shouldn’t.

  145. Ben McMillan says:

    I think that there are useful lessons to learn from the countries that have most effectively suppressed COVID, and the majority of the useful actions were not extremely draconian. Not all of the countries that are doing well are Asian, and not all the Asian countries with effective measures were China. e.g. Australia and Germany and South Korea are doing a lot better than the UK.

    1) Early and deep lockdowns were effective.
    2) Automated contact and infection tracing are helpful.
    3) Quarantine for incomers is necessary once you have cases down to a low level.
    4) Sufficient and rapid testing is useful.
    5) You need plenty of PPE.

  146. Dave_Geologist says:

    Joshua, most large droplets go less than six feet, and in cases like the Gesundheit-II machine are directional. Less so in windy conditions, but then they’d be dispersed. If they spread in still air, inverse-square law.

    A facemask won’t protect you from touching a surface on which droplets have fallen: that’s what handwashing is for, and not touching your face or your eyes. And the jury is out on whether having a mask makes you more or less likely to touch your face.

    There’s a view that viral load is a factor in the high number of healthcare workers being infected. More incidents of potential infection, more viral load. Alternatively, there is an x% chance of a single incident breaching your defences, and a person who is exposed to orders of magnitude more incidents will have a much greater risk of infection. And that’s before you consider that I shouldn’t be getting within two metres of anyone, and medical or social care requires people to get a lot closer than that so the risk will be much higher per infected person. Factor in that I spend 10-15 minutes in the shop versus 40, 50 or 60 hours working in a risky environment, and the rest is arithmetic. Of course one visit could kill me. Just as one cigarette could kill me. But I’m at much less risk than a 20-a-day man.

    I appreciate the value of a mask in stopping me sneezing on someone else. And I speak as someone who has partially controlled hayfever. But ATM there are social-care staff and other key workers who can’t get enough surgical masks, let alone N95 masks, and who don’t have the time or facilities to wash and re-use them even if they are reusable. Also, based on the Asian studies, the asymptomatic ratio is small, so at this stage in the epidemic the number of infected people a mask would protect others from is low, and one handwashing lapse can negate the mask.

    I have some masks (and plenty of toilet roll, I saw the writing on the wall when most people hadn’t seen the wall), but am saving them for when I’m sneezy or it becomes formally or informally mandatory. Scarves or home-made ones are another matter, but give that the second Cochrane study weakened the benefit found by the first and cautioned about the strength of the evidence available, I’d be even more cautious about home-made ones deployed by amateurs.

    And, as per the thread topic, Dr. Zoe is a public-health nutritionist not a physician or epidemiologist, nor a PPE or human-behaviour expert. I’ll weight her views against the consensus the same way I weight those of dabbling statisticians or retired businessmen with 50-year-old maths degrees when it comes to AGW. Particularly when the evidence is equivocal and you also have to take into account behavioural factors (riskier behaviour) and the impacts on already stretched supply and distribution chains.

  147. Joshua says:

    oilman –

    >But no, there’s no data showing that you need masks in public.

    To break that down a bit. My sense of the data/evidence is that it isn’t particularly conclusive in either direction. There is data in both directions, but it’s of dubious quality to begin with, and it gets even more unsatisfactory when you try to apply it specifically to COVID-19, in particular contexts, and with particular types of masks.

    > If you want to wear one, go ahead.

    Sure, the precautionary principle might apply there. It might help so go ahead and do it. And might also apply in the other direction, Wearing masks have potential downsides, so to be safe you should avoid doing so. And again, much of the application of that principle would depend on context and on type of mask we’re talking about.

    > Otherwise, there’s no need to get all passionate about it. (Me included. I’m being a bit of a dip.)

    Agreed.

    > I do find the similarities in argumentation to global warming interesting. It seems that some of us are flipped in perspective and trying to be the experts without anything to back it up.

    No doubt.

  148. Joshua says:

    Ben –

    > e.g. Australia and Germany and South Korea are doing a lot better than the UK.

  149. Joshua says:

    oops….

    If you look at this clip, you’ll see some interesting discussion of use of masks in the Czech Republic, Mongollia, Taiwan, HK

    I don’t mean to suggest that his points are conclusive, merely that the are probably worth consideration.

    https://youtu.be/hVEVve-3QeM via @YouTube

  150. Joshua says:

    Dave –

    > Joshua, most large droplets go less than six feet, and in cases like the Gesundheit-II machine are directional. Less so in windy conditions, but then they’d be dispersed. If they spread in still air, inverse-square law.

    Sure. But there’s the question of the aerosolized “droplets” are smaller and whether masks may have some effect of reducing transmission (both directions) of those droplets, and in particular which kinds of masks might be effective against those droplets. Also at question is whether those smaller droplets are infectious. Or whether there is a distance limit beyond which they might travel but aren’t effectively infectius.

    > A facemask won’t protect you from touching a surface on which droplets have fallen: that’s what handwashing is for, and not touching your face or your eyes.

    Of course. But they might help protect you if you’ve touched a surface with droplets and then reach towards your face only to, instead, land on the outside of your mask instead. Or they might help protect you because in wearing a mask, you’re less likely to touch your face. That might be a general principle, or it could be highly idiosyncratic, by individual

    > And the jury is out on whether having a mask makes you more or less likely to touch your face. More incidents of potential infection, more viral load.

    Right.

    > There’s a view that viral load is a factor in the high number of healthcare workers being infected.

    Just as a point of discussion, as I understand it (look at the Newscientist article I linked above), viral load can refer to the amount of virus being shedded, and not the magnitude of the infectious particles has been exposed to.

    > Alternatively, there is an x% chance of a single incident breaching your defences, and a person who is exposed to orders of magnitude more incidents will have a much greater risk of infection.

    Yes, that means that more exposure = more risk, and why healthcare workers are at higher risk – but it doesn’t speak to the question I’ve been asking you about – which is the question of why you’re saying that masks should be worn by the general public.

    > And that’s before you consider that I shouldn’t be getting within two metres of anyone, and medical or social care requires people to get a lot closer than that so the risk will be much higher per infected person. Factor in that I spend 10-15 minutes in the shop versus 40, 50 or 60 hours working in a risky environment, and the rest is arithmetic. Of course one visit could kill me. Just as one cigarette could kill me. But I’m at much less risk than a 20-a-day man.

    Again, see above. I don’t see how this information (none of which is new to me) answers my question.

    I appreciate the value of a mask in stopping me sneezing on someone else. And I speak as someone who has partially controlled hayfever. But ATM there are social-care staff and other key workers who can’t get enough surgical masks, let alone N95 masks, and who don’t have the time or facilities to wash and re-use them even if they are reusable. Also, based on the Asian studies, the asymptomatic ratio is small, so at this stage in the epidemic the number of infected people a mask would protect others from is low, and one handwashing lapse can negate the mask.

    > Scarves or home-made ones are another matter, but give that the second Cochrane study weakened the benefit found by the first and cautioned about the strength of the evidence available, I’d be even more cautious about home-made ones deployed by amateurs.

    I agree that caution is entirely appropriate here. But caution also suggests that careful use of masks may, actually, be of some value for shifting the probabilities. Whether they might shift the propabilities enough to really make a difference…or whether the different direction of probabilities (more touching face from using a mask versus some catching of aerosolized or non aerosolized droplets) lean in one direction rather than the other…..

    > And, as per the thread topic, Dr. Zoe is a public-health nutritionist not a physician or epidemiologist, nor a PPE or human-behaviour expert. I’ll weight her views against the consensus the same way I weight those of dabbling statisticians or retired businessmen with 50-year-old maths degrees when it comes to AGW.

    No doubt. I feel similarly. But nonetheless, I found her references to the meta-analysis to be useful information.

    > Particularly when the evidence is equivocal and you also have to take into account behavioural factors (riskier behaviour) and the impacts on already stretched supply and distribution chains

    No doubt. I’d still appreciated an answer to my question, which I still don’t think I’ve gotten.

    https://andthentheresphysics.wordpress.com/2020/04/03/stay-in-your-own-lane/#comment-174020

  151. Joshua says:

    Ben –

    I thought that video discussed the Czech Republic and mask use but is seems not – I must have heard about it from somewhere else. But a simple Google search will dig it up.

  152. John Hartz says:

    All masks are not created equal!

    “Hi ho Silver away!” Who was that masked man?

  153. Joshua says:

    Yeah – read that this morning JH. The difference in approach is striking.

  154. Ben McMillan says:

    It is clearly important for health care workers to have PPE. I’ve got no opinion on whether the public should, but certainly not at the expense of health care workers: I’ve given my favorite pair of woodworking goggles to a hospital worker to help keep them safe. Avoiding the kind of situation where you would need PPE is the best approach.

    John: yep, NZ is also doing well. They took much quicker action than Oz. Both these countries have a pretty serious approach to quarantine and imported pests.

  155. John Hartz says:

    anoilman: Well said. The irony of the ongoing discussion about how best to present the spread of OVID-19 is the tile of the OP, Stay in your own lane?

  156. JCH says:

    New Zealand has adopted the China strategy, which is eradication.

  157. JCH says:

    They float an article about the Chinese reopening wet markets. In the article the proof is a painted sign that has a picture of a bat for sale and a tears old video of a very attractive Asian girl eating what is allegedly bat soup. She has a very large bat in her chopsticks and appears to trying to gnaw off a part of it. This would be like somebody stabbing a whole chicken in Grandma’s soup. The article has no evidence of reopened wet markets at all, and contains several lies.

    Another planted article says 5,000 urns prove China hid 40,000 deaths. A little math and some arithmetic and some digging on Google quickly completely destroys the story. It was planted. By? Guess who?

    Crematoriums smoking night and day. Burning the hidden bodies. How freaking gullible are you people? They have very high IQ’s, and would burn the hidden bodies at a coal plant!

    On CNN Hoppy and Jim come on TV on a recent morning and start talking about startling new news, hidden by China, that SARS-CoV-2 can be transmitted by asymptomatic carriers, so I immediately start linking to a CNN article from January 26 on their twitter accounts where Fauci discusses the Chinese information that there was pre-symptomatic spread in China. Within 30 minutes Dr. Gupta and Anderson Cooper come on and correct their horrendous defaming of China. I linked to their own freaking article.

    And the BBC mistrusts the Chinese? Are western people on acid? Are they eating mushroom?. Weapons of Mass Destruction; Yellow Cake; Weapons-Grade Powder; Gulf of Tonkin. Pompeo – we lie; we cheat; we steal. He meant it. One of the few times the guy has ever told the truth.

    Trump and his little merry band of lairs are desperate to blame somebody. Don’t trust one single word coming out of the western media about SARS-CoV-2 and China, and Asia in general. They’re playing the “pit China against its neighbors” game.

    So now the little merry band of Trump liars has moved on to the W.H.O.

  158. anoilman says:

    Pro tip everyone: Don’t use disinfectant wipes on your laptop keyboard.

  159. Joshua says:

    The travesty of the US response, in a graphic form:

    https://digg.com/video/coronavirus-spread-animation

  160. Joshua says:

    Oh wait. That wasn’t bad…sonkwts look at the update:

    https://www.bbc.com/news/world-51235105

    Scroll down a bit.

  161. Dave_Geologist says:

    Joshua, you’ll have to tell me which bit I didn’t answer. I thought I’d answered them all. I’ve also seen viral load used for the number of virus particles in your nose when they start multiplying. That’s the sense in which I used it. If it’s 50 per incident, I might have 50 after a shop visit, and a nurse or homecare worker 5,000 after a shift, given the same protection.

    Not an additional reply but an expansion. As I understand it (thread topic notwithstanding) the aerosols are virus particles left after the tiny droplets have evaporated. How infectious they are, and for how long in what environments is not something I’ve seen tested. Similarly with the long-distance sneezes and the people who were still shedding virus after developing antibodies and clearing symptoms. It would require trying to infect laboratory animals (ferrets appear to be the animal of choice), which I appreciate takes longer and requires ethics clearance which would hopefully be accelerated. The PCR test, again as I understand it, just detects viral RNA. It doesn’t test whether the lipid shell remains intact and the spikes are still active. Just like a DNA test doesn’t tell you an Egyptian mummy is still alive. The cat study inferred aerosol infection between cats in different cages, but I’ve seen cautionary comments that if you have a handful of cats squeezed into each other’s personal space, even in separate cages, they won’t just be breathing normally but hissing and spitting at each other and spraying urine to scent-mark.

    Fundamentally I don’t think it’s a big deal risk-wise if you follow the other measures, at least at this stage when numbers infected are low and for most it would be a placebo. I’m more concerned about the supply chain in countries where surgical masks are not routinely used by the public and are in short supply for those at greatest risk, and risky practices in societies where citizens haven’t had years of practice using them to keep pollution at bay. I worry from the brouhaha about them that suddenly switching and making them compulsory will leave millions of people desperate for a magic bullet thinking they’re Superman and acting like Typhoid Mary.

  162. Joshua says:

    Dave –

    > Joshua, you’ll have to tell me which bit I didn’t answer.

    You said above:

    > Because healthcare workers don’t have the privilege of staying 2m away from their patients and not spending more than 10-15 minutes in a room with their patients? Or indeed staying 2m away from each other and not spending more than 10-15 minutes in a room with each other?

    Which I interpreted to basically be saying, isn’t it obvious why the following:

    > Zoe Harcombe: “I am not aware of any debate about the value of masks among healthcare workers – only among the general public. Yet why would masks protect them and not us?”

    Is not a well thought out question.

    I interpreted that to mean that you were saying it’s obvious that healthcare workers not being able to limit contact or maintain a 6′ distance, obviously, explains why healthcare workers benefiting from masks doesn’t imply that non-healthcare workers would likewise benefit from masks.

    Here’s my question about that. I don’t see a solid mechanistic explanation why either of those factors would obviously explain why the benefit of wearing masks for healthcare workers (given they can’t meet those criteria) necessarily implies nothing about the benefit of mask-wearing for the general public. The simple fact of being able to (usually) maintain a 6′ distance or limit exposure does not, IMO, seem to imply that there is a binary relationship between the benefit of mask-wearing to heatlhcare workers vs. non-healthcare workers.

    > I’ve also seen viral load used for the number of virus particles in your nose when they start multiplying. That’s the sense in which I used it. If it’s 50 per incident, I might have 50 after a shop visit, and a nurse or homecare worker 5,000 after a shift, given the same protection.

  163. Joshua says:

    Damn it…

    > I’ve also seen viral load used for the number of virus particles in your nose when they start multiplying.

    Yes, I’ve seen both also. But I think that the distinction is a worthwhile one to consider when discussing these issues.

    > As I understand it (thread topic notwithstanding) the aerosols are virus particles left after the tiny droplets have evaporated.

    Perhaps. My understanding is somewhat different in that there isn’t really a clear distinction in that sense. IOW, there are a variety of “in-between” states – related to the size of the particles and the amount of moisture that has/hasn’t evaporated. My understanding is that there many not really be a binary aereosol/non-aerosol taxonomy.

    > How infectious they are, and for how long in what environments is not something I’ve seen tested.

    My understanding is that there is a lot of uncertainty there. Which is part of the reason I suspect that non-healthcare workers may, and I stress the uncertainty there, may benefit from mask wearing.

    I’ve read some people say that if the virus is infectious through aerosols, there there is no benefit to wearing non-professional level masks, as they can’t necessarily filter out the smaller particles that would be aerosolized. There are two reasons I question that determination. The first is that I think there is not a clear size delimiter for what is/isn’t considered an aerorolized particle. The second is that I don’t understand why, even if a particle is so small that it could conceivable pass through a non-professional level mask, it would seem to me that some portion of particles that small would nonetheless get caught up in some of the material of a non-professional level mask. In other words, a non-professional mask might not be 95% effective, but it might be 30% effective – which might mean that it is worthwhile wearing as long as the potential demerits of wearing a mask (i.e., increased likelihood of fact touching, increased contact spread through improper touching of mask when taking on and off, increased potential of risks taken under a sense of security, etc.) don’t outweigh that benefit of a 30% reduction in the aerosolized particles.

    > I’m more concerned about the supply chain in countries where surgical masks are not routinely used by the public and are in short supply for those at greatest risk,

    No doubt. But that doesn’t really apply to the question of whether people should be wearing home-made masks when out in public.

    > and risky practices in societies where citizens haven’t had years of practice using them to keep pollution at bay.

    Also, I don’t really know how to wear the potential costs from improper use vs. potential benefits from improper and proper use. So that is why I questioned why you seemed to reject the potential of a net benefit to the general public wearing (home-made in particular) masks.

    > I worry from the brouhaha about them that suddenly switching and making them compulsory will leave millions of people desperate for a magic bullet thinking they’re Superman and acting like Typhoid Mary.

    Adding an element of compelling people to wear masks isn’t something that pertains to my original question. As far as I see that, there’s a whole ‘nother level of complication there. Not entirely dissimilar to the kinds of thorny questions related to governments mandating the use of various surveillance techniques to combat the virus.

  164. JCH says:

    The Chinese have everybody wear a mask. The results are obvious. Their reasoning apparently is not. Of course not, they’re commies. Can’t lose sight of that.

    Their almost universal use of masks doesn’t have to be perfect in order to be critically important; they don’t even have to be close to perfect. One, their essential workforce went out and did their essential work with great deal of confidence; two, if strict quarantine denied access as a viral host at a level of 70%, masks did not have to do one heck of a lot to raise it to 85%, at which point you likely have a proxy for herd immunity; new cases plummet to almost zero, active cases quickly drop close to effective zero. And that is exactly what happened.

    Their scientists and doctors had a successful SARS strategy; we don’t.

  165. izen says:

    @-JCH
    “two, if strict quarantine denied access as a viral host at a level of 70%, masks did not have to do one heck of a lot to raise it to 85%”

    Given the rate of infection among healthcare workers who are using good masks with proper training, and strong motivation to be careful, I think it very unlikely that masks given anything like a 15% improvement in your chances of avoiding infection.

    The main reason for the enthusiasm for wearing masks, or touting them as some sort of significant protection is to create a feeling of control and autonomy that is almost entirely false.
    Faced with a situation where there is no choice a person can make that has an effect, people look for a ritual that gives the fake reassurance of meaningful action. Apart from a communal response of isolation/quarantine we are completely at the mercy of random chance. That has always been the case with a new infectious disease, it is how SARS and MERS were controlled. The psychological necessity is to find some option that gives a false sense of being able to change a impersonal and arbitrary destiny.
    Whether it is mask wearing, ventilator availability, or most toxic, blaming China.

  166. Joshua says:

    izen –

    > Given the rate of infection among healthcare workers who are using good masks with proper training, and strong motivation to be careful, I think it very unlikely that masks given anything like a 15% improvement in your chances of avoiding infection.

    Pehaps. But given the greater risk for Healthcare workers by likely orders of magnitude compared to people walking around in a supermarket, we’d expect that a mask usage could make a huge huge difference and still result in high rates of infection among Healthcare workers – particularly the shortage of the highest quality PPE and the overwhelmed status of many institutions. I don’t really see a diet link between the high rates of infection with Healthcare workers and the likelihood of masks making a difference for the general public. The conditions are so different that making a link seems problematic to me. It’s kind of like aggregating mortality rates across all age ranges. Imo, rates for 90nyear olds should be considered independently from rates for 20 year olds.

    > The main reason for the enthusiasm for wearing masks, or touting them as some sort of significant protection is to create a feeling of control and autonomy that is almost entirely false.

    That’s certainly a factor… Abut “the main reason?” FWIW, that seems to me like an over-confident statement. There are quite a few highly qualified experts who think there could be a significant benefit from the general public wearing masks, fully independently of your speculated “main reason.”

    > Faced with a situation where there is no choice a person can make that has an effect, people look for a ritual that gives the fake reassurance of meaningful action.

    This is not mere ritual for fake reassurance. These are questions of probability and possibility in a condition of high levels of uncertainty. Of course grasping for a relief from that uncertainty is a factor… but rhe uncertainty exists nonetheless. People grasping for a facile answer doesn’t make the root uncertainty just disappear.

    > Apart from a communal response of isolation/quarantine we are completely at the mercy of random chance.

    Completely? Well, I can’t agree. Sure, random chance is the predpminsting variable beyond isolation, but I think “complete” is too strong. We don’t know to what extent other behaviors may contribute.

    > or most toxic, blaming China.

    I think that blaming China deserves its own category and doesn’t quite fit with the potential of mask-wearing to make a (small?) difference.

  167. Joshua says:

    Maybe it’s just me, but I don’t get how so many people are so confident about so many things.

  168. Joshua, showing doubt seen as a sign of weakness?

  169. Joshua says:

    Anders –

    Surely, that’s part of it.

    Hmmm

    OK, maybe that’s part of it.

  170. JCH says:

    izen, that is ridiculous. Their essential workers are not typically in presence of very sick people in a closed space. They were keeping the lights on, keeping the food delivered, etc. I do not know how often they got fresh masks, but it is likely way more often than our healthcare workers in hotspots like NYC as China makes 50% of the world’s masks.

    The resistance to Asian superiority here just astounding.Our State Department and national security folks are producing wet marketing/rhino-horn/funeral urn lies at the speed of light to keep the western public from seeing it.

    They stopped an outbreak of a very contagious viral respiratory disease in its tracks, and all people can talk about is: South Korea.

    We F’ed it up.

    They killed SARS-CoV-2 to a population essentially little different than zero. Can you cite a single example of western medicine/public health accomplishing anything remotely like that? What’s the population of SARS-CoV-2 in the western world? Moving up too fast to bother counting. Probably 15 million bodies worldwide, except China, that are brim full of the stuff.

  171. izen says:

    @-JCH
    “They killed SARS-CoV-2 to a population essentially little different than zero. Can you cite a single example of western medicine/public health accomplishing anything remotely like that?”

    No, but then no other nation enacted a full lock-down on all travel and total quarantine for anyone who had been in the vicinity of a positive case when the total number of deaths was still less than 300.
    Healthcare workers in Wuhan seem to have been just as susceptible as they are in other places, although accurate data is hard to find.
    I am not discounting the role of masks, they are certainly of some benefit if worn by a infected contagious individual.
    But the main method that the Chinese used to great effect is the same as was used against the black death in the 1300s. Full quarantine with tough restrictions on any movement except for the bare minimum of key workers for ~70 days.
    Even now the lifting of those restrictions is only for those with a mandatory smartphone application that shows they are healthy and have not been in recent contact with anyone confirmed to have the virus. That requires a means for the State to location track everyone who is allowed out, and comprehensive testing to detect >80% of the positive cases.
    A slight improvement on painting a cross on the door.

    I see no Western society with either the means or intent to follow that methodology.
    Many Western societies are not even prosecuting a sufficiently strict quarantine, and without the testing and location tracing relaxing the measures they have imposed is most likely to cause a second peak.
    With or without mask wearing.

  172. John Hartz says:

    Here’s a dimension of the COVId-19 pandemic in the US that is more in our collective lane of experience than is much of the medical discussion occurring on this thread…

    After the fossil fuel industry spent hundreds of millions of dollars undermining climate science, it’s easy to see how epidemiology came next.

    Decades of Science Denial Related to Climate Change Has Led to Denial of the Coronavirus Pandemic by Neela Banerjee & David Hasemyer, InsideClimate News, Apr 8, 2020

  173. John Hartz says:

    izen: You wrote:

    But the main method that the Chinese used to great effect is the same as was used against the black death in the 1300s. Full quarantine with tough restrictions on any movement except for the bare minimum of key workers for ~70 days.

    Even now the lifting of those restrictions is only for those with a mandatory smartphone application that shows they are healthy and have not been in recent contact with anyone confirmed to have the virus. That requires a means for the State to location track everyone who is allowed out, and comprehensive testing to detect >80% of the positive cases.

    What is the source of this information?

  174. Ben McMillan says:

    In the UK, a high proportion of healthcare workers working with COVID patients, taking various stringent measures that they are trained for (washing hands several times per patient) including wearing masks (although the focus on this seems excessive/misguided) are now getting ill (sorry, this is anecdotal, I don’t have a authoritative source). By far the best protection is not going near other people.

    Agree with Izen that the main measure that has been effective in China and elsewhere has been very stringent lockdown and case tracing. If you can’t leave your house, masks are a moot point.

    But many Western countries and non-Chinese Asian countries are taking effective measures against Covid and new cases appear to be reducing. It isn’t as black and white as ‘China or bust’. e.g. South Korea has partial lockdown measures in place, and much closer to normalcy than any Western country. Australia, New Zealand appear to be doing well. Germany is doing much better than many other EU countries.

    I’m not convinced that tracking apps would be unacceptable in the West: so many ‘impossible’ things have happened over the last few weeks. It is a bit of an eye-opener as to what is possible when societies actually take a problem seriously.

  175. Dave_Geologist says:

    Joshua, izen’s point about range and frequency of contact is a better one than my mechanistic attempt. Given that the complaint from nurses and doctors (but not social-care workers) is not that they have nothing, but that they have surgical masks and aprons not N95 masks and gowns, if repeated close-range exposure is not more risky than occasional distant exposure, why are healthcare workers, skilled in hygiene and PPE-wearing, being infected at such rates, IIRC higher in many countries than Joe Public?

    I too found her references to the meta-analysis to be useful information, because one of the nice things about a meta-analysis is that you can use it as a guide to whether noisy data is being misconstrued as evidence (a Type 1 Error). One way is to stratify the studies by (objectively) good to poor, either methodologically or by sample size. If you find the better studies support the hypothesis more strongly, you get a warm feeling. If the reverse, you don’t. Similarly, if you have two studies a decade apart, the second adding new studies to the first, and the second supports the hypothesis more strongly than the first, you get a warm feeling. When, as here, the reverse is true, you don’t.

    At least for me, this particular sub-thread has run its course.

  176. Why the lane of standard methodical peer-review does not exist in these times:

    View at Medium.com

    “The ethical imperatives in terms of knowledge building are for doctors to convey the conditions and results of their actions as precisely as possible. It is only after the battle that the necessary sorting work will take place. Right now medical teams around the world are trying things. They communicate their experiences with each other through new means of communication: through social networks, using new means of communication (various messaging). Hundreds of articles have been published on a disease that did not exist three months ago. For example on BiorXivwhere the work has not been validated before publication, as is the case in scientific journals. Are these works useless because they have not received the seal of peer accreditation? No, you just have to take them for what they are: partial information on which you have to exercise a critical eye. Peer validation has not disappeared; it applies after publication. It’s a new way of collaborating that didn’t exist before because the act of publishing was expensive: you had to print physical books, distribute them, etc. The cost of withdrawal in the event of an error was high, it is essentially zero today (we can update the information). This way of validating new knowledge, which goes through width rather than depth, is faster and takes more effort. You have to exercise critical thinking all the time, but it is also the promise of faster progress than we have ever known. It must also encourage reflection, more than reflexes, ethics because new questions arise.”

  177. anoilman says:

    Interestingly, regions of high air pollution have higher mortality rates from COVID-19. This study shows that small (1 mg/m3) increases in long-term PM 2.5 air pollution are associated with a 15% increase in coronavirus deaths. This links coronavirus, air pollution, and climate change in a very unflattering light.
    https://projects.iq.harvard.edu/covid-pm

  178. Dave_Geologist says:

    This is well worth a read, IMO: Jennifer Nuzzo: We’re Definitely Not Overreacting to COVID-19.

    It covers a lot more than reaction – transmission, differences from other viruses, comparison with flu, sobering realism on how soon we could deploy a vaccine globally…

  179. Joshua says:

    Dave –

    > if repeated close-range exposure is not more risky than occasional distant exposure,

    I never intended anything that I wrote to be interpreted in such a way as to suggest that repeated close-range exposure is not more risky than occasional distant exposure.

    Of course I agree that repeated close-range exposure (and longer duration of exposure) would be more risky than occasional distant exposure. But that, in itself, doesn’t imply that the public wearing masks, even non-medical grade masks, would have zero net beneficial effect (or a net negative effect given the likelihood of improper usage).

    What I’m saying is that from what I’ve seen, there’s a lot of uncertainty w/r/t the public wearing non-medical grade masks. In that case, it seems to me that the precautionary principle may apply. Of course, that could run in either direction – as it could suggest that untrained people should use non-medical grade masks when out in public.

    But even there, the question is complicated in that the question should be broken down w/r/t the potential to prevent infection of the mask-wearer, versus helping to prevent greater spread of the virus because people sneezing and coughing and are wearing masks that might prevent larger droplets from landing on surfaces likely to be touched by others.

    > One way is to stratify the studies by (objectively) good to poor, either methodologically or by sample size.

    Yes, I really love it when people break down meta-analyses by criteria such as power, CI’s, and other factors reflecting on the quality of the research being surveyed. We need more of such meta-analysis.

    > At least for me, this particular sub-thread has run its course.

    Fair enough.

  180. John Hartz says:

    Another downside to chilling out on weed to wile away the time…

    If you’re smoking weed to ease your stress during the coronavirus pandemic, experts say it’s time to think twice.

    Smoking marijuana, even occasionally, can increase your risk for more severe complications from Covid-19, the disease caused by the novel coronavirus.

    “What happens to your airways when you smoke cannabis is that it causes some degree of inflammation, very similar to bronchitis, very similar to the type of inflammation that cigarette smoking can cause,” said pulmonologist Dr. Albert Rizzo, chief medical officer for the American Lung Association. “Now you have some airway inflammation and you get an infection on top of it. So, yes, your chance of getting more complications is there.”

    Smoking weed and coronavirus: Even occasional use raises risk of Covid-19 complications by Sandee LaMotte, CNN, Apr 10, 2020

    https://www.cnn.com/2020/04/10/health/smoking-weed-coronavirus-wellness/index.html

  181. John Hartz says:

    Speaking of testing for COVID-19…

    Critics say the Nordic island country should also have closed all of its schools and stopped tourism. But the government says it believes that new cases have peaked.

    Iceland’s ‘Test Everyone’ Goal Has Skeptics, but It May Be Working</strong< by Aimee Ortiz, Europe, New York Times, Apr 9, 2020

  182. John Hartz says:

    anoilman: You wrote:

    This links coronavirus, air pollution, and climate change in a very unflattering light.

    Are you implying that ecoronavirus, air pollution, and climate change can be linked in a flattering lightunder certain circumstances? 🙂

  183. John Hartz says:

    Dave_Geologist : The Jennifer Nuzzo article is spot on. Thank you for bringing it to our attention.

    PS – You can bet your sweet bippy that the Pretend President of the US and hs minions will not read it. He supposedly relies heavily on Dr Oz for medical advice.

  184. Everett F Sargent says:



    RoW = Rest of World = World – (EU+US+CN+IR) = most exponential to date (or linear in log-normal space)

  185. Joshua says:

    Anders –

    Here’s a good article in the uncertainty (touches on armchair epidemiology)

    https://nymag.com/intelligencer/2020/04/best-case-scenario-for-coronavirus.html

    My brother pointed out to keep in mind

    If you are looking to find something that happens 1 case in 1000, Assume you always find it when it is there. Assume that your false negative rate is 1%. Then in 1000 cases you find 1 true positive and 10 false positives, so the probability that you have a true positive given a positive test is <10 %

    And then there's developing a reliable test, and then getting it out (who would have predicted such crap with the testing so far?), then verifying someone's results.

  186. John Hartz says:

    Everett F Sargent; Given how suspect much of the underlying data is, are the graphics you posted meaningful?

  187. Steven Mosher says:

    ○ One of the cases confirmed on 8 April has been found to be linked to the bar Liquid Soul in Seocho-gu, Seoul. In total, 5 cases have been confirmed from Liquid Soul since 6 April.

    ○ From the wine bar UnWined in Pyeongtaek City, Gyeonggi Province, 1 additional case has been confirmed, bringing the total to 18 confirmed cases (wine bar = 14; family/acquaintances of confirmed cases = 4). Further epidemiological investigation is underway.

    ○ From Gyeongbuk Province, during the epidemiological investigation (and testing) on a new case, 3 family members and 1 co-worker have been found confirmed with COVID-19. Further investigation into chains of transmission and contacts is underway. (The figures on Table 2 are based on cases reported to KCDC before 0:00 of 10 April and may differ from above.)

  188. Steven Mosher says:

    “There’s tons of science on this. The experts are 100% glaring at, and looking through, and working on this, and have been for oh, 100 years or more.

    Zero citations.
    Zero data.

    Again.

    No data on the relevant claims.
    Zero.

  189. Joshua says:

    Steven –

    In my town, cases have recently gone up. I called my congressman to find out if there has been ANY contact tracing. I’d like to know if there’s been community spread or if the new cases are from known contacts.

    I’ll be shocked if I even hear back, let alone get answers to my questions. As per that article I just linked, I can visualize a way out if this, but we are distressingly far from heading on that path.

    This is insanity.

    And the level of facile/politically “motivated” reasoning in the “skept-o-sphere” outstrips even my expectations.

    Only going to contribute to making it that much harder.

  190. Everett F Sargent says:

    JH,

    “Given how suspect much of the underlying data is, are the graphics you posted meaningful?”

    It is all that we have to date, the most reliable are likely to be death statistics. If one goes about casting aspersions or arguments from incredulity, from whatever ideology, particularly conspiracy theories, we all are left with nothing.

    But sure, let us all talk nonsense about COVID-19 by not using any numbers at all and call it scientism. /:

  191. Everett F Sargent says:

    Here are a couple more graphs …

    x-axis = confirmed daily cases
    y-axis = daily deaths

    And yes, you can complain about the 8-9% slopes and that the y-intercepts are greater then zero (deaths occur for zero confirmed).

    The correlations appear to be robust. And since daily deaths lag daily confirmed by seven days, I’ll make the following forecast for close of business next Friday …

    EU+USA = 134k cumulative deaths (from 88.6k).
    World = 155k cumulative deaths (from 102.5K).

  192. anoilman says:

    Mosher: Look.. I know its not your field, but that doesn’t mean no one at all on the planet isn’t and never has looked at your present concern. It means you lack the required knowledge, experience and understanding.

    Don’t freak out. Should I say it again?
    Keep Your Distance,
    Wash Your Hands,
    Don’t Touch your face, and,
    If it makes you feel better wear a mask?

    FYI… Willard chastised me for pooping on your education.. so I don’t go there. Intuitively, what you say makes sense to me. But if you ever did any technical work, then you’d know that the answer doesn’t come from your intuition, there is the other 99% perspiration required to get a result.

  193. Dave_Geologist says:

    And a TED Talk with embedded transcript and associated summary article:

    The quest for the coronavirus vaccine

    COVID-19 needs a Manhattan Project

  194. Ben McMillan says:

    EFS: you may be better off with a 7-day average given the very strong ‘weekend-effect’ in some countries. Also, divisions on the y-log scale are odd when the leading digit on the axis tic label is 2.

  195. Ben McMillan says:

    (but just quibbles, these are useful/interesting graphs)

  196. JCH says:

    Ad5-nCoVn – Adenovirus type 5 vector that expresses S protein – CanSino Biologicals

    They’ve started phase 2.

  197. John Hartz says:

    Everett F Sargent: When we discuss climate science data and graphics, we always refer to uncertainty and the quality and consistency of the data. The same should apply to COVD-19.

    I also question your assertion: … the most reliable are likely to be death statistics.

    According to news reports, for example, Italy only reports deaths that occur in hospitals. If so, this excludes people who have died elsewhere due to complications from COVID-19 infections.

    As far as I can tell from the media reports I have seen and read, each country “does its own thing with respect to measuring infections and deaths. This suggests to me that both metrics aggregated globally are on the low side.
    ..
    FWIW: I do not appreciate the lecturing included in your response to my question.

    PS: The next time you post graphics, please include a link to the source(s).

  198. John Hartz says:

    ATTP: Suggested fodder for a new OP…

    Carbon pricing is often presented as the primary policy approach to address climate change. We challenge this position and offer “sustainability transition policy” (STP) as an alternative. Carbon pricing has weaknesses with regard to five central dimensions: 1) problem framing and solution orientation, 2) policy priorities, 3) innovation approach, 4) contextual considerations, and 5) politics. In order to address the urgency of climate change and to achieve deep decarbonization, climate policy responses need to move beyond market failure reasoning and focus on fundamental changes in existing sociotechnical systems such as energy, mobility, food, and industrial production. The core principles of STP can help tackle this challenge.

    Opinion: Why carbon pricing is not sufficient to mitigate climate change—and how “sustainability transition policy” can help by Daniel Rosenbloom, Jochen Markard, Frank W. Geels & Lea Fuenfschilling, PNAS Journal, Apr 8, 2020

    https://www.pnas.org/content/early/2020/04/07/2004093117

  199. Ben McMillan says:

    John, for what it is worth, although the death data is not perfect, most commentary suggests it is substantially better than the ‘case number’ for understanding the progress of the epidemic. Probably only 10% of infections end up coming to health system attention (Iceland data) although this is very variable. And the number diagnosed is not even particularly consistent as health resources get overwhelmed or testing improves.

    The proportion of deaths to cases in Germany vs UK, for example, is probably very different mostly because of better testing/tracing in Germany (e.g. look at how many have been tested).

    This suggests that the official figures might underestimate the death toll (in Lombardy) by a factor 2.5 but this is very uncertain:
    https://towardsdatascience.com/covid-19-excess-mortality-figures-in-italy-d9640f411691

    So the official death numbers are not great, but more accurate and probably more consistent than the case numbers (for looking at epidemic progress).

  200. John Hartz says:

    Ben McMillan: I concur. I also expect that there will be a lot of post-mortem adjustments to the data as time go by. Lots of Phd dissertations may also be written about this for specific countries.

    Do you happen to know whether or not the WHO has in place protocols for countries to follow in collecting and reporting pandemic statistics? If there are, I suspect that WHO does not have any meaningful way to enforce compliance with the protocols.

  201. Ben McMillan says:

    No, sorry, don’t know about that. My impression is that WHO isn’t resourced well enough to deal with something like coronavirus so ‘enforcement’ of data standards spread across ~200 countries is unlikely to happen given other priorities (budget $4 billion a year). Just dealing with the last ebola outbreak was probably a stretch (imagine how much a comparable military operation would cost).

  202. Phil says:

    In case people haven’t seen this, I thought I’d just drop it here;
    https://cmr.asm.org/content/cmr/20/4/660.full.pdf

    The concluding remarks (on p24 of download, p683 of journal)
    Coronaviruses are well known to undergo genetic recombination (375), which may lead to new genotypes and outbreaks. The presence of a large reservoir of SARS-CoV-like viruses in horseshoe bats, together with the culture of eating exotic mammals in southern China, is a time bomb. The possibility of the reemergence of SARS and other novel viruses from animals or laboratories and therefore the need for preparedness should not be ignored.

    This was published in the wake of SARS in 2007.

  203. David B Benson says:

    According to Worldometer, the number of new cases in the USA has stopped rising, day by day.

  204. Phil says:

    Ben McMillan
    Probably only 10% of infections end up coming to health system attention (Iceland data) although this is very variable.

    The Kings College London research https://covid.joinzoe.com/post/covid-isolation gives an estimate of 1.9 million people in UK with symptomatic COVID on April 1st. Against this, the cumulative confirmed cases to 11th April was 70,000. Those figures would mean 3.7% of infections coming to health system attention in the UK as an upper bound i.e. it assumes the Kings College snapshot estimate for the population as a whole on 1st April is same as the cumulative total to April 4th (allowing a week of self isolation before admittance to hospital). However the latter is bound to be greater than the former, especially since the Kings College research also shows that a nett 0.4 million people recovered between 1st-5th April

  205. John Hartz says:

    Everett F Sargent: Thanks for the source links. As I understand it, WHO is the source of the Johns Hopkins data.

  206. anoilman says:

    WHO has protocols and definitions in place to allow the collecting of comparable statistics. Most nations and regions attempt to adhere to them. However, not every country follows them, and there may be intervening policies. Such as in one region they may only seek to find a single positive per household rather than testing each member.

    It really isn’t possible to get consistent measurements. There is a lot of variance in local compliance.

  207. John Hartz says:

    Another thought-provoking analysis by David Wallace-Wells…

    The Best-Case Scenario for Coronavirus Is That It’s Way More Infectious Than We Think by David Wallace-Wells, Intelligencer, New York Magazine, Apr 10, 2020

    https://nymag.com/intelligencer/2020/04/best-case-scenario-for-coronavirus.html

  208. John Hartz says:

    Joshua: No harm, no foul.

  209. izen says:

    Here is someone not just out of their lane, but apparently way off-road…

    “This is a very brilliant enemy. You know, it’s a brilliant enemy. They develop drugs like the antibiotics. You see it. Antibiotics used to solve every problem. Now one of the biggest problems the world has is the germ has gotten so brilliant that the antibiotic can’t keep up with it.
    And they’re constantly trying to come up with a new – people go to a hospital and they catch – they go for a heart operation – that’s no problem, but they end up dying from – from problems. You know the problems I’m talking about. There’s a whole genius to it. We’re fighting – not only is it hidden, but it’s very smart. Okay? It’s invisible and it’s hidden, but it’s – it’s very smart.”

  210. Ben McMillan says:

    That intelligencer article has some serious flaws, actually: too many things said, and not careful or quantitative enough. The idea that mortality predictions are wrong because of undetected cases seems wrong/confused. A lot of it seems to be based on his own reading of a bunch of tweets and he needed to run the whole lot by a relevant expert.

    Consider:
    “A new, small-sample serological survey in Germany holds more promising results: 14 percent of those tested carried coronavirus antibodies, meaning they’d already been exposed the disease, orders of magnitude more than is suggested by their confirmed case count of less than one-tenth of one percent of their population.”

    Actually, this is one of the worst hit towns in the whole of Germany: the number of cases of Coronavirus is 0.15% of the population in Germany, and 0.6% in this town (as of the date of this article):
    https://www.thelocal.de/20200402/how-german-scientists-hope-to-find-answers-on-coronavirus-in-countrys-worst-hit-spot

    The German study this is based on

    Click to access zwischenergebnis_covid19_case_study_gangelt.pdf


    (site is down?) suggests an infection fatality rate of 0.37% according to the Spiegel article:
    https://www.spiegel.de/wissenschaft/medizin/corona-in-heinsberg-jeder-siebte-koennte-immun-sein-a-14bd9e0e-0c7e-4775-a8b0-1611ececd870

    And this is not that different to other estimates of the infection fatality rate (IFR). Most certainly not ‘orders of magnitude’ lower. It is quite a lot lower than usual figures for the case fatality rate (CFR) of a few percent. Because these are two separate things.

    Mostly the reason the models are getting things ‘wrong’ is the R0 value (number of people a single infection gets passed on to) is highly uncertain, especially once lockdowns are in place, and the numbers depend very strongly on when lockdowns are put into place. Not anything to do with the CFR or IFR.

  211. Ben McMillan says:

    Phil: yes, results from that app indicate that the proportion of cases coming to attention of the UK health system is currently very much lower that 10%. Also, seems consistent with the very high case fatality ratio in the UK. The number of ‘cases’ doesn’t tell you much about the number of infections.

  212. Dave_Geologist says:

    John, I’ve only briefly skimmed the Wallace-Wells piece but AFACS it suffers from the same blind spot as the wishful-thinking interpretation of the Oxford study (which didn’t claim 100 or 1000 asymptomatic cases for every symptomatic – just put it out as a hypothetical and said, based on death or hospitalisation data alone, we can’t disprove it as yet).

    Trouble is, that’s like saying based on UAH alone we see only weak surface warming, loosely coupled if at all to anthropogenic CO2 emissions, and if it is coupled ECS must be at or below the bottom of the IPCC range.

    It flies in the face of all the data from countries which enforced widespread testing and quarantine, where everyone with a fever was tested, and where most had colds or flu or other non-Covid infections, even at the peak of the epidemic. Even if you think everything coming out of China is a lie, you have other SE Asian countries and captive populations on cruise ships. The consensus is about half are asymptomatic, maybe a bit more, maybe a bit less, maybe 80% or so at most which is the Lewis equivalent for ECS. At one extreme of the range, but not quite outside it.

    For us to achieve rapid herd immunity without killing hundreds of thousands in the UK and millions in the USA that number needs to be 99% or 99.9%, not 30%, 50% or even 90%. We should soon have the answer to that in the UK. We’re out of the flu season here, so the prediction would be that it is already much more widespread than flu and close to 100% of symptomatic NHS staff tested should be positive, along with asymptomatic contacts from before we hit the millions-infected mark. Actually we probably don’t have to wait. Just crunch the numbers from Italy and Spain. Or don’t ignore 90% of the available evidence in favour of the 10% that says we can’t prove it’s not as mild as flu.

  213. Ben McMillan says:

    Even if the infection fatality rate is 0.4%, that’s still about a million dead in the US to get to herd immunity (but actually the health system would collapse so it would be worse). So the ‘reassuring message’ Wallace-Wells is selling seems to be missing the target.

  214. Dave_Geologist says:

    Re. the King’s College app: it tracks self-reported symptoms, not identified Covid-19 AFAICS. It doesn’t tell us who has something else, and doesn’t tell us who is infected and infectious but reports in as asymptomatic. If everyone who thinks they have symptoms has Covid-19, testing NHS and key workers with symptoms is a waste of time and resources. They all have it and will all test positive.

    The cold and flu season is waning, but the hay fever season is on the up and there are other sources of respiratory infections. My hay fever goes into my throat when it’s bad (whether directly or because stuff running down from my eyes and nose carries bacteria and gives me strep throat, although I had Pollinex treatment decades ago which largely suppresses that). I can tell because I get a tingling in my nose, then my tongue, then a sore throat, even though other symptoms are suppressed. Others who get a runny nose first or only a runny nose and sneezing? IIRC only 3% of Covid-19-positive have runny noses, but the premise of a large asymptomatic group is that they are not coming in and being tested. During the cold/flu/hayfever season millions of people have runny noses or sneezing but no sore throat. They’re being told not to report to the NHS as Covid-symptom-positive. In the large-numbers-asymptomatic case, most of them probably have Covid-19 and runny nose is a symptom.

  215. Ben McMillan says:

    Dave: I don’t know how the King College app works (e.g. does it somehow cross-correlate with formal diagnosis) but I’ll note that if there really are 2 million people in the UK infected (on April 1st), 10000 dead to date is not exactly reassuring (given the lag of ~2 weeks).

  216. Phil says:

    Dave_Geologist

    I don’t think your first statement is true. In their blog post about loss of smell, they explicitly say that they do actually test a subset of people who are using the app, and they report a percentage of people reporting COVID-19 symptoms who test negative for the virus. I *assume* therefore that the “machine learning” bit of this research is to identify patterns in *tested* users reporting (both positive and negative), which is then extrapolated over the untested users.

  217. Phil says:

    This post https://covid.joinzoe.com/post/how-we-are-predicting-covid-in-the-uk gives more details about how they are predicting incidence of COVID-19 from the reported symptoms

  218. Dave_Geologist says:

    Thanks Phil. I did look at their site (and had earlier in the week when someone messaged me about it), but found it hard to get details on their methodology.

    But as per my earlier comment and Ben’s, the consensus asymptomatic ratio needs to be wrong by orders of magnitude to make a difference in decision-making at this stage. Tweaks won’t cut it.

    See BMJ expert comments on the Oxford preprint:

    Ferguson said, “We’ve been analysing data from a number of Italian villages at the epicentre for the last few weeks where they did a viral swab on absolutely everybody in the village at different stages of the outbreak. And we can compare that with official case numbers, and those data all point to the fact that we are nowhere near the Gupta [the Oxford analysis] scenario in terms of the extent of the infection.”

    James Naismith, director of the Rosalind Franklin Institute at the University of Oxford, backed the call for widespread testing. “This will be necessary to test the paper’s hypothesis. The need for, the science behind, and plans to implement such serological testing are accepted and moving forward across the globe,” he said. He warned that this would “take time.”

    He added, “At this moment, nothing in [this] paper calls for or could be used to justify any change in current policy; that is, unless we all follow the current government advice on social distancing, the UK will see many thousands of deaths that could have been avoided.”

    Mark Woolhouse, professor of infectious disease epidemiology at the University of Edinburgh, said that the possibility that large numbers of people have already been infected but remain asymptomatic was “hypothesis rather than fact.”

    He said that, if it were correct, “that would not change current policy in the UK, which is focused on reducing the short term impact of the epidemic on the NHS.” It would, however, “change enormously our long term expectations” and suggests that the threat of covid-19 worldwide would diminish in coming months.

    To act as if half the population had already been infected in March would be like quitting your job and booking the holiday of a lifetime the day you buy a lottery ticket, not the day you win the lottery.

  219. Phil says:

    Ben McMillan:
    but I’ll note that if there really are 2 million people in the UK infected (on April 1st), 10000 dead to date is not exactly reassuring (given the lag of ~2 weeks).

    If there were 1.9 million people infected on April 1st (rather than cumulatively), then surely you compare that to the death count on a single day, no ? If you take a 2 week lag then yesterdays figure of 917 deaths is the best we have (11 days). Using that, and expanding that to the entire UK population of 66m gives 917 x 66/1.9 = 32 k deaths for the entire population. Or have I done something wrong ?

  220. Phil says:

    Dave_Geologist
    But as per my earlier comment and Ben’s, the consensus asymptomatic ratio needs to be wrong by orders of magnitude to make a difference in decision-making at this stage. Tweaks won’t cut it.
    Yes, I concur. I wasn’t suggesting that the Kings figures enabled a different policy approach, only that they are probably more reliable than they first seem.

    I should make clear perhaps that in my previous comment (currently in moderation), I was only attempting a rough, order-of-magnitude calculation …

  221. Ben McMillan says:

    “The latest figures estimate that 1.4 million people in the UK aged 20-69 have symptomatic COVID, a fall from 1.9 million on the 1st April”: current cases, not new cases that day. People are symptomatic on many days, but normally die only on one (even if it is Easter).

    This would possibly be a better starting point but the resulting estimate probably has order-or-magnitude error-bars each way:

    I’ll note that the KCL people are being circumspect and not weighing in on mortality estimates.

  222. JCH says:

    If they are truly asymptomatic, do they acquire immunity? How? If they don’t, how does policy adapt to that?

  223. Joshua says:

    Ben –

    > the idea that mortality predictions are wrong because of undetected cases seems wrong/confused.

    I read the article as exploring the various important parameters, that bwknfnkne of them. I didn’t read it as being remotely conclusive as to the size of the denominator, but exploring what it might mean if it is significantly bigger.

    Do you think it’s wrong/confused to consider the analyses that say the denominator is quite a. It bigger?

    I will say I thought it was a bit too optimistic about how well people are following distancing guidelines – although in my observations people are getting better (say staying 6 feet away in the supermarket) I think there’s still a long way to go and there’s a good possibility that there will be a fatigue pint where people start getting more lax again.

  224. Joshua says:

    Ben –

    > Actually, this is one of the worst hit towns in the whole of Germany: the number of cases of Coronavirus is 0.15% of the population in Germany, and 0.6% in this town (as of the date of this article):

    But in the other side, the percentage of the population tested is much higher than in the whole of Germany.

  225. Joshua says:

    Dave –

    > It flies in the face of all the data from countries which enforced widespread testing and quarantine, where everyone with a fever was tested, and where most had colds or flu or other non-Covid infections, even at the peak of the epidemic.

    Even testing everyone with a fever doesn’t really address the question. It will only be settled when a large enough random and representative sample is tested.

  226. Ben McMillan says:

    Joshua: I’d feel more like I was having a productive discussion if you would acknowledge that W-W has made a large arithmetic error. And that the German study provides evidence against W-W’s suggested very low fatality ratio.

  227. Joshua says:

    Ben –

    I saw him as stressing uncertainty and being equivocal about everything. I’m actually not following with what his large mathematical error was?

    I have seen many people reference that German study reporting the 14% rate. Are you saying that the study is wrong (from an error) ir that people are mischaracterizing the study?

    Yes, it would be a mistake to extrapolate from one town only. But again, I thought that Welles put in caveats about that. And on the other hand, we certainly should be careful about extrapolating rates from largely untested populations or populations where only people with symptoms are tested. There is some value in looking at the numbers from a towm where a large random sample was tested – even if extrapolating from that is tricky at best.

    Basically, I think that pointing to the uncertainty about the denominator, saying that the potential of it being larger is possibly a hopeful parameter, is a useful thing to do. I get the sense you disagree?

  228. Ben McMillan says:

    I read W-W as saying the number of cases in that village is <0.1%. But actually it is 0.6%.

    The point Dave and I have been making is that a 'large denominator' is only helpful in the short term (or explains issues with models) if it is huge, so a substantial fraction of people have already had the disease, and herd immunity is about to happen. But there is strong evidence that this is not the case.

    Including the Heinsberg study that W-W refers to. This is particularly annoying because W-W has claimed that it supports his argument, but the opposite is true. Contra W-W, '50 percent of infected Icelanders don’t know they are carrying COVID-19.' also suggests that this denominator is not very big.

    Also, W-W is conflating 'infectiousness' with 'proportion infected' but these mean very different things. Evidence that R0 is high has little to do with evidence for a large number of hidden infections but W-W is effectively claiming they are the same thing.

    My complaint is not so much about the denominator per-se but that W-W is sloppy and not well-informed. Piece needed more input from relevant scientists.

  229. Steven Mosher says:

    you guys may have missed this study before it was “retracted”

    The government is now strictly controlling all science on covid

    https://www.dailymail.co.uk/news/article-8094933/How-one-man-spread-coronavirus-NINE-people-bus.html

  230. John Hartz says:

    Ben McMillan: You and other commenters have convinced me not to take everything that David Wallace-Wells writes at face value. I’m probably too impressed by his narrative style than I ought to be in order to read his stuff with a critical eye. The fog of a the global pandemic we are experiencing is very much like the fog of war.

  231. Everett F Sargent says:


    I’ll make the following forecast for close of business next Friday …
    USA = 52.5k cumulative deaths (from 37k).
    EU = 122.5k cumulative deaths (from 96K).
    EU+USA = 175k cumulative deaths (from 133k).
    World = 206k cumulative deaths (from 154k).


    log-normal y-axis goes 2,4,6,8,10,12,14,16,18,20,40,60, …

  232. John Hartz says:

    Then there’s this…

    Far more people may have been infected with Covid-19 than have been confirmed by health officials in Santa Clara County, California, according to a study released Friday in preprint.

    The study used an antibody blood test to estimate how many people had been infected with Covid-19 in the past. Other tests, like those performed with nasal swabs or saliva, test for the virus’ genetic material, which does not persist long after recovery, as antibodies do.

    “We found that there are many, many unidentified cases of people having Covid infection that were never identified with it with a virus test,” said Dr. Jay Bhattacharya, a professor of medicine at Stanford University and one of the paper’s authors. “It’s consistent with findings from around the world that this disease, this epidemic is further along than we thought.”

    The study estimated that 2.49% to 4.16% of people in Santa Clara Country had been infected with Covid-19 by April 1. This represents between 48,000 and 81,000 people, which is 50 to 85 times what county officials recorded by that date: 956 confirmed cases.

    Far more people may have been infected by coronavirus in one California county, study estimates by Michael Nedelman, CNN, Apr 17, 2020

    https://www.cnn.com/2020/04/17/health/santa-clara-coronavirus-infections-study/index.html

  233. Ben McMillan says:

    John: glad you found my comment (and those of others) interesting. People are talking about that Santa Clara article over in the other thread.

    EFS: I figured that’s what the scale must be… 7 day smooth seems to slightly remove the lumpiness but I guess a weekend effect superimposed on an ‘exponential’ is not totally removed by a 7-day smooth.

  234. John Hartz says:

    Ben: The commenter herd tends to graze on the most recent comment thread regardless of the new OP’s topic. I haven’t been following the chatter on it.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.