Coronavirus - Modelling Aspects Only

[supremetwo]

It's certainly spiked near Malvern:-

https://www.bbc.co.uk/news/uk-england-h ... r-53381802
It employs about 200 seasonal pickers and packers from Eastern Europe to help with the crop.
---------------------------------------
I understand although a few of the workers had symptoms, most were asymptomatic and the positive results of 73 people came as a surprise to them.

Tests before entry into the UK needed.

[redsturgeon]

https://www.theguardian.com/world/2020/ ... y-suggests

This suggests that we should not go all in on herd immunity hopes.

John

[johnhemming]

What it says is that the numbers of antibodies go down. What it does not say is that people lose resistance to the disease.

[Mike4]

The really important question no-one seems to be focusing on is HOW MANY are asymptomatic but testing positive. Pretty high from the comment in the article but no figures given.

Here we have another (broadly) closed community which could add light to the issue of what proportion of people catch it and are infected but never develop symptoms, which is massively helpful when trying to guess stuff like HIT. Let's hope the scientists are all over this like a rash....

[servodude]

What it says is that the numbers of antibodies go down. What it does not say is that people lose resistance to the disease.

What it said was:
"the findings suggested people could become reinfected in seasonal waves and that vaccines may not protect them for long".
- that does to my reading sound as though they are suggesting that resistance is lost
-sd

[johnhemming]

https://www.medrxiv.org/content/10.1101 ... 20148429v1

Antibody (Ab) responses to SARS-CoV-2 can be detected in most infected individuals 10-15 days following the onset of COVID-19 symptoms. However, due to the recent emergence of this virus in the human population it is not yet known how long these Ab responses will be maintained or whether they will provide protection from re-infection. Using sequential serum samples collected up to 94 days post onset of symptoms (POS) from 65 RT-qPCR confirmed SARS-CoV-2-infected individuals, we show seroconversion in >95% of cases and neutralizing antibody (nAb) responses when sampled beyond 8 days POS. We demonstrate that the magnitude of the nAb response is dependent upon the disease severity, but this does not affect the kinetics of the nAb response. Declining nAb titres were observed during the follow up period. Whilst some individuals with high peak ID50 (>10,000) maintained titres >1,000 at >60 days POS, some with lower peak ID50 had titres approaching baseline within the follow up period. A similar decline in nAb titres was also observed in a cohort of seropositive healthcare workers from Guy′s and St Thomas′ Hospitals. We suggest that this transient nAb response is a feature shared by both a SARS-CoV-2 infection that causes low disease severity and the circulating seasonal coronaviruses that are associated with common colds. This study has important implications when considering widespread serological testing, Ab protection against re-infection with SARS-CoV-2 and the durability of vaccine protection.

There is evidence that people obtain resistance to SARS-Cov-2/Covid-19 from infections by other coronaviruses. It would be surprising if that form of T Lymphocyte resistance was not also created by an infection from SARS-Cov-2. Similarly we do not internationally know of anyone who has had a serious SARS-Cov-2 infection twice.

Whereas I would accept the point that we don't know how long infection resistance would last and that if there is a vaccination it probably needs boosters at some frequency. The report does not say that people become immediately prone to re-infection. The nub of the resistance to Coronaviruses is that it does fade, but people retain some resistance it is not all lost. (perhaps I should have used the word "all" previously)

[johnhemming]

https://www.medrxiv.org/content/10.1101 ... 20154294v1

Abstract
It is widely believed that the herd immunity threshold (HIT) required to prevent a resurgence of SARS-CoV-2 is in excess of 50% for any epidemiological setting. Here, we demonstrate that HIT may be greatly reduced if a fraction of the population is unable to transmit the virus due to innate resistance or cross-protection from exposure to seasonal coronaviruses. The drop in HIT is proportional to the fraction of the population resistant only when that fraction is effectively segregated from the general population; however, when mixing is random, the drop in HIT is more precipitous. Significant reductions in expected mortality can also be observed in settings where a fraction of the population is resistant to infection. These results help to explain the large degree of regional variation observed in seroprevalence and cumulative deaths and suggest that sufficient herd-immunity may already be in place to substantially mitigate a potential second wave.

[johnhemming]

Having noticed the source of UK data on hospital admissions we can, of course, get a peak date for admissions
This gives England
https://coronavirus-staging.data.gov.uk ... me=England

Other nations or NHS regions can be obtained from there.
The peak admission dates are
England 31st March
Scotland 1st April
NI 2nd April (two peaks)
Wales 1st April

Generally admissions to hospital occur 2 weeks after infection (some say 13 days).

The deaths figures on the same website are based upon date of report rather than date of death.

[Mike4]

Generally admissions to hospital occur 2 weeks after infection (some say 13 days).

Does that include the (average) five day symptom-free incubation period?

From the anecdotes I read, patients manage fine at home for about a week after symptoms appear, then they improve and think they are on the mend. At which point they either turn out to be right or regress into serious illness. Patients typically seek hospital admission (if they are going to at all) on day 8 to day 12 after becoming symptomatic AIUI.

[johnhemming]

From infection so if you take 5 days to get symptoms the 8-12 for admission you are talking about 13-17 days for admission which is the same sort of period.

In any event it substantiates the argument that the lockdown did not constrain the peak of infections. The Welsh Admission figures look quite different to the other national ones. I would take it that Wales was further from herd immunity than the other nations which is why the number of admissions to hospital did not go down as quickly as they did elsewhere. There could be other reasons, however.

[swill453]

The Welsh Admission figures look quite different to the other national ones. I would take it that Wales was further from herd immunity than the other nations which is why the number of admissions to hospital did not go down as quickly as they did elsewhere.

I don't think anywhere in the world has been shown to have herd immunity, has it?

Scott.

[johnhemming]

It depends what you mean by "herd immunity". If you mean the level of immunity that lasts like measles for decades then no. However, if you mean the level of resistance to infection which in itself holds back the expansion of infection then that is where the argument is.

Different people have different views on this.

cf
https://lemonfool.co.uk/viewtopic.php?f ... 60#p326896

[zico]

Latest ONS excess deaths data report - England and Wales only. It's for week 27 (week ending 3rd July).

https://www.ons.gov.uk/peoplepopulation ... g3july2020

Between weeks 13-27 (From start of epidemic in England & Wales) there were 58,830 excess deaths.

Looks like the first wave has definitely ended now. Third week in a row that excess deaths were (slightly) below the 5-year average. There are still 500-600 Covid-19 deaths recorded, so this is more than matched by 600-700 fewer non-Covid deaths recorded. Quite a big difference between hospital deaths (going down) and private homes deaths (going up).

In Week 27, the number of deaths registered was 0.5% below the five-year average (43 deaths fewer), this is the third consecutive week that deaths have been below the five-year average; the numbers of deaths in care homes and hospitals were also fewer than the five-year average (88 and 634 deaths lower respectively), while the number of deaths in private homes was 755 higher than the five-year average.

The decline in non Covid deaths is likeliest to be due to people having died earlier from Covid. There have been 58,830 excess deaths, and using the figures estimated elsewhere than people dying from Covid on average die 10 years earlier than they would otherwise have done, we'd expect 5,883 fewer deaths per year, which is about 115 per week.

However the non-Covid excess deaths is about 600-700 per week, so there must be other factors - though this is getting into speculation.
One likely factor is that because everyone is being more hygenic and having less contact with others, there's likely to be a drop in infectious diseases. An enforced slower pace of life under lockdown and restrictions is likely to reduce stress which may also account for fewer deaths generally.
Another possibility is that people aren't going to hospitals for surgery with a possibility of a fatal outcome. There are 27,000 deaths per year from surgery, which equates to 520 deaths per week, although not all surgery will have stopped during the pandemic, so the difference will be less than 520.

It's also been estimated that 7,000 extra people per year could die of cancer with a worst-case scenario of 35,000 if cancer services don't get back to normal - though it's not clear over what time period this would be (see BBC link below for what scanty details are available). Deaths due to non-treatment of cancers will show up eventually in excess deaths stats, though these may take several months to work through the system.

https://www.bbc.co.uk/news/health-53300784

[servodude]

What's the reasoning behind not treating cancer ?
Reallocation of resources or risk mitigation or something else.

I would hope that it can be brought back quickly and pragmatically.

I know that the hospitals I deal with in Aus have ceased sleep tests due to the risks of infection through CPAP, but oncologists have been treating my sister in law for cancer throughout.
(And my hypochondriac second cousin who was convinced he had contacted COVID in Feb on a Pacific cruise has spent much of the time since being treating in hospital for Guillaume Barre syndrome - he was chuffed to be actually sick for once! )
-sd

[johnhemming]

I have put this here as it is more appropriate than the thread it is on which is considering the ONS estimates for the number of people getting Covid-19
https://lemonfool.co.uk/viewtopic.php?f ... 00#p328701

Another way of estimating this is to look at hospital admissions and assume that these are people who caught the disease 2 weeks earlier. Hence we cannot directly compare the week being discussed.

However, we can say that in the week 13-19 July 641 people were admitted to hospital. The difficulty, of course, is knowing what proportion of people who catch the disease end up in hospital. If it is 10% then 6,410 people caught the disease if it is 1% then 64,100 caught the disease.

I have had a quick glance at the ONS site to try to find the result from 2 weeks ago, but for the same week they have a range between 10,500 and 38,500 people catching the disease with a 95% confidence interval with their preferred figure as being 19,600.

The peak admissions figure was about 3000 and peak deaths over 1000 so perhaps we can estimate from what I think is a high IFR of 0.5% that 1.5% of people who catch the disease (including asymptomatic) hence the infection rate two weeks ago would be 42,700.

What this does say is that the ONS figures are in the same sort of ball park as working backwards from hospital admissions although I think the IFR is too high. At a rough guess in 2 weeks the number of infections goes down by about 40% (to about 60% of the starting value). Hence one would conclude the infection rate in the last week would be about 25,000.

If we then take those figures and apply them to the peak infection rate one would conclude a figure of around 1.5 Million people in a week and something like 10 Million in England as a whole.

As far as I can tell this ignores, however, people who are infected in care homes and die there. (I know the ONS figures ignore people infected in care homes).

[feder1]

Using the worldometer,s data, I have been plotting UK Deaths in comparison with Cases as specified 14 days earlier and divided by an eyeball factor of 6.

I.e, dividing the latest cases by 6 would predict deaths in a fortnight,s time.

However, lately, this crude correlation has deviated and the dividing factor seems to be 10.

Now when I plot all the data using factor 10, it looks as though at the peak of deaths, a much bigger proportion of cases were expiring, 1/6 instead of 1/10.

This suggests that the hospital treatment is a lot better now, maybe medically or becauses the staff are less overworked.

Yesterday,s cases predict 75 deaths in a fortnight.s time. Still too high a daily death rate since the 7 day moving average is stubbornly refusing to fall.

[dealtn]

Yesterday,s cases predict 75 deaths in a fortnight.s time. Still too high a daily death rate since the 7 day moving average is stubbornly refusing to fall.

Purely out of interest does anyone know what the daily death rate for influenza would be in the UK at this time of year (in a "normal" year)?

[Nimrod103]

Using the worldometer,s data, I have been plotting UK Deaths in comparison with Cases as specified 14 days earlier and divided by an eyeball factor of 6.

I.e, dividing the latest cases by 6 would predict deaths in a fortnight,s time.

However, lately, this crude correlation has deviated and the dividing factor seems to be 10.

Now when I plot all the data using factor 10, it looks as though at the peak of deaths, a much bigger proportion of cases were expiring, 1/6 instead of 1/10.

This suggests that the hospital treatment is a lot better now, maybe medically or becauses the staff are less overworked.

Yesterday,s cases predict 75 deaths in a fortnight.s time. Still too high a daily death rate since the 7 day moving average is stubbornly refusing to fall.

All the press reports seem to indicate that British doctors have learnt a great deal about treating this disease, and reducing mortality, with dexamethasone etc, and perhaps less use of ventilation. However, there may be other explanations, such as the virus cull has scythed through the most vulnerable in the population (perhaps more related to genetic susceptibility), and is now naturally dying out.

[zico]

Yesterday,s cases predict 75 deaths in a fortnight.s time. Still too high a daily death rate since the 7 day moving average is stubbornly refusing to fall.

Purely out of interest does anyone know what the daily death rate for influenza would be in the UK at this time of year (in a "normal" year)?

It's in the latest ONS coronavirus report.
5-year average for "Deaths where pneumonia is mentioned" is around 1,400 per week (so 200 per day) over the last few weeks.
Currently 2020 average for a similar time period is around 1,000 per week (so about 30 per day fewer than normal).
Latest weekly ONS figures for Covid-19 registered deaths are 606,532,366 for last 3 weeks respectively.

https://www.ons.gov.uk/peoplepopulation ... 10july2020

Also link below to interesting report showing causes of all deaths in the UK by type. Pneumonia is quite a small category compared to heart disease and Alzheimer's.

https://www.ons.gov.uk/peoplepopulation ... 2001to2018

[Mike4]

All the press reports seem to indicate that British doctors have learnt a great deal about treating this disease, and reducing mortality, with dexamethasone etc, and perhaps less use of ventilation. However, there may be other explanations, such as the virus cull has scythed through the most vulnerable in the population (perhaps more related to genetic susceptibility), and is now naturally dying out.

One thing the press seem silent about which supports this view, is the forecast impact from the Bournemouth beach scrums and BLM marches/riots ought to be showing up by now in the infection stats I think. But they aren't AFAICS.