Advertisement



  1. Things you need to know about the new ‘Conversations’ PM system:

    a) DO NOT REPLY TO THE NOTIFICATION EMAIL! I get them, not the intended recipient. I get a lot of them and I do not want them! It is just a notification, log into the site and reply from there.

    b) To delete old conversations use the ‘Leave conversation’ option. This is just delete by another name.
    Dismiss Notice

Coronavirus - the new strain XVIII

Discussion in 'off topic' started by matt j, Apr 16, 2021.

  1. jimbob75

    jimbob75 pfm Member

    Of course. That's why I'm not using those stats to try and make a point.
     
    hc25036 likes this.
  2. gavreid

    gavreid pfm Member

    That was the first national breakdown of deaths from Delta under the criteria for reporting that have been used throughout most of the pandemic - the numbers are still small but are statistically relevant.

    "the 28-day case fatality rate for Delta remains low (0.1%), though mortality is a lagged indicator and the vast majority of cases are still within the 28 days of follow-up required." [see below]

    Here's more on the PHE technical briefing this morning - the risk assessment for Delta has been updated too.

    Now Public Health England (PHE) has revealed that up to 96% of new Covid cases in the UK involve the Delta variant, and that it has about a 60% increased risk of household transmission compared with the Alpha variant.

    https://www.theguardian.com/world/2021/jun/11/delta-variant-is-linked-to-90-of-covid-cases-in-uk
     
  3. gavreid

    gavreid pfm Member

    That is precisely what you are trying to do, of course. The poor folks probably got run over by a car, had asthma at the same time with a touch of aids for good measure - the thing is I've heard it all before and it's just how the early deaths were explained away last March and again in September.
     
  4. Seanm

    Seanm pfm Member

    Fair to say though that without context those figures are mood music.
     
  5. gavreid

    gavreid pfm Member

    The thing to look out for is how long it takes 40 odd deaths to become 90 odd and then 180 odd...
     
  6. John Phillips

    John Phillips pfm Member

    For contrast, in May in UK there were 77,731 cases (all variants) and 246 deaths (all variants). If I have the right figures (from Daily summary | Coronavirus in the UK (data.gov.uk)), that's 0.3% death rate. That's down from April when it was 0.8%; further down from March when it was 1.7%; and further down from 3.9% in January and February.

    EDIT: in June so far the death rate is 0.1%.

    We seem to be going in the right direction but we don't know for sure what the impact of Delta will be. Even if it is worse than before, the calculation in the tweet seems to not offer that clarity.
     
    Sue Pertwee-Tyr and deej92 like this.
  7. gavreid

    gavreid pfm Member

    I know I said i'd wait to update the hospitalisations figures but I'm not being drawn into an argument about deaths on these numbers. The PHE risk assessment now says the following:

    "Iterated analysis continues to show vaccine effectiveness against Delta is higher after 2 doses but that there is a reduction for Delta compared to Alpha. There is uncertainty around the magnitude of the change in vaccine effectiveness after 2 doses of Oxford-AstraZeneca vaccine." [my emphasis]

    Below we see, in all likelihood, that hospitalisations are beginning to grow exponentially (i.e. a straight line on a log scale) and the yellow line represents a doubling time of 20 days just to guide the eye until there is enough data to fit sensibly. It's still early days, which is what I have been trying to explain - we just can't expect these to be running away yet. We are, however, going entirely in the wrong direction and hoping that deaths will not follow suit - they're much too low to follow day by day at present.

    [​IMG]
     
  8. Andrew C!

    Andrew C! Been around a while....

    Moving round the ground - masks on. When seated, masks off.
     
  9. gavreid

    gavreid pfm Member

    Sorry can't agree with the logic - understand it might stop the beer flowing though ;) Be careful tomorrow won't you, my brother was going today but decided against yesterday, he just had his 2nd dose this week. The game seems to be quite nicely set up if England bowl well this aft.
     
  10. Andrew C!

    Andrew C! Been around a while....

    Ta. I’ve had both doses(AZ) some time ago, and will comply with mask wearing. My view is I’ve been in enough primary schools in the last 8 weeks and not caught anything (no mask wearing in class, just communal areas - schools policy btw)…
     
  11. gavreid

    gavreid pfm Member

    Indie Sage - theme, should we vaccinate children?



    Admissions of children to ICU is higher as a proportion than for seasonal flu.
     
  12. Woodface

    Woodface pfm Member

    Interestingly there have been no recent infections at my daughters school when it was quite a thing earlier in the year.

    Personally I would like the Govt to put it back by two weeks to buy some time for more jabs.
     
    John Phillips likes this.
  13. gavreid

    gavreid pfm Member

    A couple of weeks will make little difference - sorry.
     
  14. John Phillips

    John Phillips pfm Member

    I am beginning to feel that is sensible. Assuming cases are increasingly less important per se and it's the undesirable consequences that matter, analysing the admissions and deaths, with the recent data I have, shows a large range of possibilities. There's too much noise in too small a data set. A week or perhaps two might make things clearer. I do acknowledge that government sources will have more comprehensive data so this caution may be misplaced.

    The recent admissions graph illustrates the issue well. It is certainly possible to draw an exponential with a troublingly small 20-day doubling time. However you get a very small regression coefficient if you do that, revealing that the curve fit is very poor. Fitting an exponential curve for the highest possible regression coefficient gives a doubling time of 53 days which is much better than 20. However the improved regression coefficient (R² = 0.37) is still too small for high confidence in the model.

    EDIT: And analysing deaths data shows even more potential variation between best and worst cases.

    Exponential growth models do not inevitably imply fast growth. But this is often the case. Herein lies the dilemma that when things could happen quickly you often need to know from too small a data set that you have to act. The data set as I see it supports neither panic nor complacency: analysis gives too wide a variation between best-case and worst-case predictions. But maybe this does indicate that more data would reduce the variation and so inform a better decision.
     
  15. gavreid

    gavreid pfm Member

    8 125 cases today (I think the doubling time is looking like 9 days or so), 17 deaths and 147 admissions (8th). The admissions figure is always back-loaded to the end of the week but the increase is undeniable and much worse than my yellow line in recent days. (I could nudge it downwards a spot and parallel but that would make matters worse.) The yellow line represents Alpha behaviour under the previous conditions just as a reminder. Vaccination will just not keep up with this rate for a very considerable time, given the time that it takes to be effective. John, I hope you're right but I think you're mistaken right now. Here's the graph with today's update

    [​IMG]
     
    tiggers likes this.
  16. mandryka

    mandryka pfm Member

    Yes I remember that from last week. Put aside the fact that they don’t define vaccine effectiveness. (i.e. effectiveness against what?) Now why would there be uncertainty about AZ but not about Pfizer, given that so much AZ has been used in the UK? Can anyone think it through? I’m just not clear enough about the logic of the situation.
     
  17. mandryka

    mandryka pfm Member

    Can you spell out what that means? there will be more admissions, but more people are being vaccinated all the time, so I guess that in a short time the rate of growth of admissions should start to decline . . . but I can’t think it through.
     
  18. gavreid

    gavreid pfm Member

    Effectiveness compared to their previous analyses.

    It's really the hare and the tortoise. The hare seems to be winning comfortably - the vaccine will hopefully catch up eventually but it's slower, much slower, so there will be another peak in cases and deaths. It's not just getting the jabs into arms but the time that it takes to boost immunity thereafter.

    The point about AZ is the key concern for me and you jumped on it - it will take time but it seems that there is a weakness in the vaccine even at double dose - it could just be that 2 doses takes longer to take effect than PZ or that AZ is waning, I really don't know.
     
    tiggers likes this.
  19. John Phillips

    John Phillips pfm Member

    My primary conclusion is that the data does not support any particular conclusion yet. I guess you have a conclusion on the pessimistic end. And I am not saying that is wrong.

    There has clearly been a significant decoupling of consequences from cases up to date. That is supported by the data. However, even that will not easily withstand a prolonged rise in cases at the recent 10-day case doubling time, so there will be another wave.

    EDIT: 9.7 days doubling time (R² = 0.82).

    The uncertain issue is the Delta variant. Looking at the data I think how big the wave will be depends very strongly on assumptions for which we have no solid basis today. I am not predicting which way things will go - just pointing out the wide uncertainly.
     
  20. gavreid

    gavreid pfm Member

    I have an open mind, I report on the data and challenge complacency
     

Share This Page





Advertisement


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice