Viewing 40 posts - 18,161 through 18,200 (of 18,584 total)
  • The Coronavirus Discussion Thread.
  • Premier Icon BillMC
    Full Member

    I’m bone fido working class, it’s a dog’s life.

    Premier Icon oldnpastit
    Full Member

    I stumbled across this on the WHO website, an estimate of IFR from seroprevalence data taken from about 80 separate studies.

    https://www.who.int/bulletin/online_first/BLT.20.265892.pdf

    Makes for interesting reading if you were wondering what the IFR is.

    Some of the studies cited have some pretty interesting numbers, including a slum in Mumbai with 54.1% testing positive for antibodies.

    EDIT: also kind of interesting given who the author is.

    Premier Icon anagallis_arvensis
    Free Member

    Son of a tradesman

    Posh!
    You arent working class anymore, like I’m not either and I have more degrees than both my parents had o levels, saying that I have no idea if my father even did them, but my step dad certainly had none!

    Premier Icon anagallis_arvensis
    Free Member

    150 deaths today, 80 a week ago!

    Premier Icon kelvin
    Full Member

    I’m not sure what data we’re waiting for to take proper action? Time for everyone to make their own call in this again… the gov only want to follow not lead… just as in the spring. We’ve been hoping they’d learnt the lesson from the spring that they need to be preemptive… where as, in fact, they’ve learnt that they can get away with waiting ‘till most people have looked at the data and made the move already. I’m aware that there’s little you can do as a teacher AA, but companies need to step back to minimal social contact, and we need to do the same in our off time as well.

    Premier Icon Northwind
    Full Member

    oldagedpredator
    Full Member

    I’d never heard of ‘uncle sugar’ until this week – I see it’s another americanism. More importing of more clap trap like defunding.

    I saw it starting to creep in with obvious US bots/posters on reddit and facebook and stuff where they’d just got the localisation wrong and thought everything was about the US. But now it seems to have caught on a bit with real people, mostly as a sort of shibboleth for easily idenfifying absolute ****.

    “Defunding” is different though, it’s a real and possibly very important thing that just has had massive amounts of effort and bots and stuff thrown at derailing and misrepresenting it.

    (Defunding started out as a call to take things like mental health, homelessness and drug responses off the police, because they keep on shooting people for the crimes of autism and poorness, and take away the funds that were spent on that with the police, and give it to more appropriate operators.

    Then, when black lives matter started really getting traction and more and more people were seeing how some police departments and police unions act like organised crime gangs, defunding became part of suggestions to break that deadlock- it wsa never about “let’s have no police”, it was about “if these particular police act above the law and above public redress, let’s defund them, and spend the money on some less murdery police instead”. The way law enforcement is funded in the US is insane, so stopping things like civil forfeiture and other self-interested negative feedback loops would be sensible.

    Premier Icon thecaptain
    Free Member

    That Ioannidis thing is the usual bollocks that’s been ripped apart on twitter by people far more knowledgeable than me. Basically, cherry-picked to give an answer as low as he can manage, and it’s still a lot higher than he was saying earlier.

    eg here

    Premier Icon seosamh77
    Full Member

    Someone posting that ioannidis study else where was what prompted my fag packet maths 4 pages ago.

    Under 70 = 6772 deaths / 3,955,877 * 100 = 0.17%

    Over 70 = 31632 deaths / 709,622 * 100 = 4.45%

    ps if you take the uk as a whole, puts it as:

    7% of 66.65m – 4.6655m

    so 43429 / 4665500 * 100 = 0.93%

    I came up with that. (happy to hear criticism, I’m just about understanding, so have no attachment to these numbers, but they are based on the assumption that half a billion people have had covid, ie 7% of the population, which is the last concolusion on the ioannidis paper)

    Tbh the biggest conclusion I came away from reading it was that, there’s probably a lot of countries not recording their covid deaths correctly?

    It’s clear that the IFR is low for under 70s, which we all know anyhow, but either countries don’t have a lot of over 70s or their reporting is wrong. I mean criticise the uk all we like, they deserve it, but I think the reporting here is at least something approaching accurate,

    disclaimer, an utter laymans view of things, would be interested to hear more opinions. My mind is pliable on the matter if someone can present interesting views.

    Premier Icon TiRed
    Full Member

    That Ioannidis thing is the usual bollocks

    Generally, it’s the quiet ones you should listen to. As time passes, the relatively early null hypothesis will come to pass. That this is a pathogen that does not have widespread pre-existing immunity, has low mortality in the young, but a considerable burden on the elderly, and that reinfection cycles look like the other endemic coronaviruses.

    I have little time for him. Wanting something to be true and bending interpretations of data to meet that desire has to have testable hypotheses if you all want to be scientific.

    Premier Icon oldagedpredator
    Full Member

    As for the public sector, we kid ourselves what governments can really do, particularly in the short term. Short of training the army to run pcr machines (Machines magically from where?), you need business. Companies that are competent and have a track record of delivery. I have no issue with that.

    Measures of competency are another thing. As is tasking the impossible. At least we seem to be following the Germans along the road to devolved testing and tracing.

    For me sustained under investment is as much about future effectiveness as it is about our capabilities right now. I have no problem with bringing in extra private sector capacity. it’s just not allying to the existing expertise for what seems like no other good reason than the dogma of we don’t need the state that frustrates me.

    We have cut public services to the bone – the inevitable ‘belt tightening’ measures are now going to involve bone shaving. Euphemistically – that’s going to be quite painful. There’s no oh we can rely on charities – they’re facing their own significant financial challenges. Many of the small ones aren’t even going to be at bone shaving it’s lights on or lights out. If the solution to cover off that shortfall becomes an outsourcing one then we are just haemorrhaging money.

    Premier Icon TiRed
    Full Member

    Completely agree. It’s the pretence that the public sector will provide and is then found seriously wanting. Germany has devolved healthcare and has done very well. We’ve been sending tests to their labs due to over-capacity. Italy was sending their ITU patients!

    Premier Icon MoreCashThanDash
    Full Member

    companies need to step back to minimal social contact, and we need to do the same in our off time as well.

    It’s heading that way – most employers seem to be working towards that again anyway. MrsMC and I are both involved with a variety of local youth groups, which as it stands, have exemption to carry on even at Tier 3. We are firmly of the view that “just because you can, doesn’t mean that you should”. She’d already pulled the plug on indoor Guide meetings before the GG announced it, and has a couple of outdoor evening ideas lined up.

    The other groups we are involved in we are on the committees, so it’s not our sole decision. We spent an hour or so in bed this morning discussing options, tactics in committee meetings, and working out funding/budgets. On our weekend lie in. Who said romance was dead? 🙂

    Premier Icon stumpyjon
    Full Member

    companies need to step back to minimal social contact

    My company have never stopped 300 office based staff sent home the week before lock down, nk expecration of return before next year. We did plan a small trial return with my team, 10 people back in 2 days a week in an office that normally holds 70, canned that idea 6 weeks ago at the first hint of a pick up in cases before we returned. Out MDs approach is why take the risk with peoples health and business viability when we’re managing work from home. We want to go back, but only when it’s sensible to do so. Boris’s pleas to return for the sake if Starbucks fell on deaf ears.

    Premier Icon dissonance
    Full Member

    Completely agree. It’s the pretence that the public sector will provide and is then found seriously wanting

    Was that the case? As I understand it the slow start with testing were primarily due to PHE and their consultants (since like any public sector institution they have had resources slashed aside from when it comes to handing it to private sector consultants) pushing a highly centralised system using a handful of private companies with even NHS trusts being discouraged from using their own labs.

    Boris’s pleas to return for the sake if Starbucks fell on deaf ears.

    Yeah the large company I worked for response could be summed up as “I am sure you read what Johnson dribbled out but dont worry we arent idiots”.
    They have opened offices and are keeping them open where possible but they have made it clear its only for those who for whatever reason find working from home difficult.

    Premier Icon reluctantjumper
    Full Member

    Quick question: How many ICU beds do we have in total? I’m watching the number of Covid cases in hospitals rise fast and just want to know at what point the whole system will be full. I know certain places have more capacity than others and that the Nightingale Hospitals will take some of the strain off the rest of the system but I can’t find a definitive answer on our capacity for regular beds.

    Premier Icon MoreCashThanDash
    Full Member

    Yeah the large company I worked for response could be summed up as “I am sure you read what Johnson dribbled out but dont worry we arent idiots”.
    They have opened offices and are keeping them open where possible but they have made it clear its only for those who for whatever reason find working from home difficult.

    As did most of the Civil Service. 😎

    Premier Icon TiRed
    Full Member

    And the good news is that there is evidence of negative curvature in the rate of growths of hospital admissions…Below is a projection NOT a model prediction – it only assumes that things will follow in the same direction, and that deaths lag admissions. It will be wrong as intervention is coming of course (Tier2/3). But you get the idea. At 3000 admissions/day you are back in April.

    Premier Icon dirtyrider
    Free Member

    anyone got any advice here

    work for the NHS (psyc rehab, in patient) and rules constantly changing,

    shift is 7am to 8pm – 1 hour break,

    covid positive staff has provided a knee jerk reaction to enforcement of rules, which has evolved to at the moment

    masks at all time, unless alone in an office with a closed door,

    masks not to be removed to eat or drink while on shift unless in an office with a closed door,

    can they actually enforce this? eating, not an issue, i can eat pre work, and on my break, but drinking? come on now, dropping my mask to take a swig of coffee whilst socially distanced, is that something they can actually enforce, considering if i ask to go to an office to have a drink, im technically taking a break

    Premier Icon thecaptain
    Free Member

    Ok I said 500 deaths by mid nov so there is clear blue water between my prediction and @TiRed above.

    I don’t look forward to being proved right.

    Premier Icon BillMC
    Full Member

    I’m hearing about new funded partnerships evolving between the NHS and university medical schools for doing t and t. It’s an amazing discovery that medical schools and the NHS seem to know a bit more about all this stuff than Deloitte and Serco.

    Premier Icon Nobeerinthefridge
    Free Member

    Dirtyrider, use a straw?.

    I’d be seeking your unions advice tbh, 13 hours is too long to work with one drinks break.

    Premier Icon TiRed
    Full Member

    I don’t look forward to being proved right.

    Indeed – sadly the negative curvature is modest at best but highly statistically significant. That’s a 90% prediction interval for daily observations from a Poisson regression. Uncertainty may be over-dispersed with a wider interval. But it’s not looking great – “Baked in”. London, the SE and the SW are about 10-14 days behind, as these bulk up the numbers, a pure log-linear exponential may be a better description – hence the 500. But London is locking down earlier than the North.

    Premier Icon ElShalimo
    Free Member

    TiRed – Does the curve have 3 StdDev ranges?

    Premier Icon yourguitarhero
    Free Member

    My work were pretty ahead of the ball – they told me not to come in to work or even do any work in March. And stopped sending me any money which must have saved them some important money.

    Premier Icon Sandwich
    Full Member

    Nightingale Hospitals will take some of the strain off the rest of the system

    Nightingales are due to be staffed from the referring hospital so that probably won’t help.

    Premier Icon TiRed
    Full Member

    3 StdDev ranges?

    No – that’s a 99% interval (0.5th to 99.5th percentiles). The plot is 5th to 95th percentiles (1.645 standard deviations), so 9/10 observations might be expected to fall in that range. Standard curves work as follows:

    99/100 inside 3 standard deviations (P0.5 to P99.5)
    19/20 inside 2 standard deviations (P2.5 to P97.5)
    9/10 inside 1.645 standard deviations (P5 to P95)
    8/10 inside 2.24 standard deviations (Inter-decile range, P10 to P80)
    1/2 inside 0.68 standard deviations (Inter-quartile range, P25 to P75)

    All assuming a normal distribution. A Poisson count does reduce to normal for large counts and assumes the standard deviation is the root of the mean. The data may be over dispersed which means more variability (called negative binomial), but I haven’t formally tested this. Predicting too wide an interval is usually a hiding to nothing due to outliers – 9/10 is reasonable to reject the model at the 10% significance level.

    Hope that makes sense. As @thecaptain says – if there are 500 deaths per day by mid Nov, one can reject the model with 90% confidence because it is outside the prediction interval. That’s how inference works.

    [tl:dr] prediction of the future must include variability. If future values fall outside of the range one can reject the model that made the prediction with some confidence. How much confidence is left to the user. Statistics works at 9/10 to 19/20 times by convention.

    Premier Icon ElShalimo
    Free Member

    Thanks. My stats is a bit rusty despite what’s happened this year

    I thought that Poisson was best used for events with very low frequency and high consequences but rarely seen eg modelling huge earthquakes.

    -ve Binomial is more commonly used for modelling higher frequency phenomena like windstorms where it tends to be more empirical (but what we’ve observed is not the best representation of what can happen)

    (My line of work is very different to yours)

    Premier Icon TiRed
    Full Member

    NHS seem to know a bit more about all this stuff than Deloitte and Serco.

    Remind me about that App from NHSX. I agree that some clinical knowledge will be helpful for judgement. But you need systems to make things work operationally. The NHS runs on XL from what I can gather. That’s what they are buying.

    Premier Icon TiRed
    Full Member

    I thought that Poisson was best used for events with very low frequency and high consequences but rarely seen eg modelling huge earthquakes.

    Plenty of options for describing count data. Poisson is the simplest and is derived from events at a constant rate per unit time. If there are too many zeros in your data, the next is called “zero-inflated” which says zero’s happen more often, and are treated separately. But if there is something to count, it behaves Poission (variance = mean). Then if the variability is a bit off (more or less than Poission), the next tool in the box is negative binomial. If there are more zeros than you want, then zero-inflated negative binomial. And so on…

    When the numbers get large, everything tends to look like a normal distribution 😉

    Premier Icon ElShalimo
    Free Member

    👍

    Premier Icon oakleymuppet
    Free Member

    Purely out of curiosity, aren’t both Binomial/Poisson distributions, the distributions of independent events? How are hospital admissions independent events?

    Application of these distributions to model variability is limited, in part, because of the need to assume independence and homogeneity. For studies of communicable disease, application may be inappropriate because of lack of independence. For example, if one person in a group develops the ‘flu’, others in that group have higher risk, reflecting dependency of disease occurrence. Insituations like these, the investigator may attempt to modify the model to account for the dependency.9

    Does your model do take that into account TiRed?

    Premier Icon TiRed
    Full Member

    Does your model do take that into account TiRed?

    To describe events such as radioactivity counts, yes, Poission has a constant rate per unit time. In modelling disease events, the underlying rate of events is time-varying (which captures the disease transmission dynamics). So the mean rate is predicted by, for example, an SEIR model, but individual observations of numbers in any day would be Poisson distributed. People tend to come in unit amounts – hence count data is described, Inconvenient, I know 😉 . Then to construct a prediction interval, either calculate in closed-form with the above assumptions, or perform bootstrap simulations and summarise across the replicates. The above plots are the summary of 10,000 realisations of the model.

    When counts get low and zeros appear, zero-inflated models come into play.

    Premier Icon oakleymuppet
    Free Member

    Cheers TiRed, I’m now going to go away and bang that into my skull and break out the google kung fu until I think I fully understand that post.

    People tend to come in unit amounts – hence count data is described, Inconvenient, I know

    😀

    Premier Icon ratherbeintobago
    Full Member

    @reluctantjumper The only number I know off the top of my head is that Wales has 189 at baseline, but, for example, most places round here increased ITU capacity by at least 50%. We’ll run out of staff before we run out of equipment or bed spaces.

    The Nightingale locally was no help last time round, as the admission criteria were very strict and as other people have said the specialist staff can’t be conjured out of thin air; those who could be redeployed from other areas had already been to service increased intensive care provision in the existing hospitals. It’s hard to see how it will be different this time.

    Also, there’s not enough capacity in the system to do a surge and continue diagnostics/treatment elsewhere, which the ‘hidden’ harm resulting from that.

    Premier Icon kelvin
    Full Member

    Which is why it’s wise to remember that we didn’t “protect the NHS” back in the spring… we all but shut down normal NHS care to divert resources to keeping Coronavirus sufferers alive. We can’t do that during the winter. Well, we can, but the damage will be far greater than in the spring.

    Premier Icon scotroutes
    Full Member

    Lots of folk now reporting a 5-day wait for test results. Todays numbers from Scotland are artificially low due to an issue with the UK testing apparatus. I suspect the other nations might have a similar problem. That’s all very handy when we’re actually trying to determine whether recently introduced measures are working and whether or not additional measures are required.

    Premier Icon reluctantjumper
    Full Member

    It is all rather convenient isn’t it. I’m just watching the numbers of patients in hospital as if that gets too high then no amount of test, track or tracing will help.

    @ratherbeintobago – that’s the only figure I could find too, nothing for England at all. I’m only asking as my dad’s chemo is being ramped up next week and the consultant said it would go ahead unless all the ITU beds were reserved for Covid19 patients and that they were close to capacity already. If they cancel the treatment then he might not make it through the winter purely due to him wasting away.

    Premier Icon TiRed
    Full Member

    Critical care beds

    Unlike most other categories of hospital bed in the NHS, the total number of critical care beds has increased in recent years. Monthly data for January 2020 indicates there were around 5,900 critical care beds, which is 13 per cent higher than the 5,200 in January 2011 (see Figure 2). Of these, 4,100 (just under 70 per cent) are for use by adults, with the remainder for children and infants.

    https://www.kingsfund.org.uk/publications/critical-care-services-nhs?utm_source=facebook&utm_medium=social

    Premier Icon kelvin
    Full Member

    the total number of critical care beds has increased in recent years

    Barely. In fact as a proportion of the population, we have fewer critical care beds now.

    Premier Icon kelvin
    Full Member

    Rustled up a graph from WHO data…

    WHO

    Top line is all EU countries, bottom line is UK only.

    Source

    A bit scary that we’re forming the bottom of the range of all countries there. World beating.

Viewing 40 posts - 18,161 through 18,200 (of 18,584 total)

You must be logged in to reply to this topic.