Coronavirus - Modelling Aspects Only

[dealtn]

Death rate = 4,034/73,000 = 5.5% in 2 months (equating to 33% per year).

Again, I'm trying not to be picky (honestly!) but you really can't do that statistically, unless you are sure you have a linearly driven probability distribution over time, and frankly that's unlikely. You can't just multiply by 6 to make that an annual rate.

What is probably more likely is that for those surviving 28 days (the measure) there is a higher risk of dying on day 29, than day 365. In the same way that we have mortally tables that show a 90 year old more likely to die "tomorrow" than a 30 year old, you will have a mortality effect where you are more likely to die of a relapse, or complication from the possible effects of treatment/surgery closer to the date of that treatment/surgery.

Statistics, particularly in "real life" are complex.

[Itsallaguess]

We have evidence from the USA which lifted restrictions early, had an initial surge in cases, and after the inevitable lag, these increased cases are now causing an increase in deaths.

The telegraph article linked to earlier presented some data that showed that the vast majority of the new cases arising in the UK hotspots are in people aged under 65, and the article explained that this is one of the reasons why we're still experiencing low daily-death rates in the UK.

That might suggest that there's still a lot of 'respect' for the virus in the older UK population generally...

It would be interesting to see if there's any data for the second US wave, in terms of the age groups getting infected this second time around, as it's not clear that the same 'respect' for the virus also exists in that older age group worldwide..

Difficult to say without seeing some hard data, of course, but the above might explain why some countries are seeing larger numbers of second-wave cases quickly drop through to their deaths charts, whilst other countries are not, in quite the same way...

Cheers,

Itsallaguess

[zico]

Death rate = 4,034/73,000 = 5.5% in 2 months (equating to 33% per year).

Again, I'm trying not to be picky (honestly!) but you really can't do that statistically, unless you are sure you have a linearly driven probability distribution over time, and frankly that's unlikely. You can't just multiply by 6 to make that an annual rate.

What is probably more likely is that for those surviving 28 days (the measure) there is a higher risk of dying on day 29, than day 365. In the same way that we have mortally tables that show a 90 year old more likely to die "tomorrow" than a 30 year old, you will have a mortality effect where you are more likely to die of a relapse, or complication from the possible effects of treatment/surgery closer to the date of that treatment/surgery.

Statistics, particularly in "real life" are complex.

Yes, you're absolutely right that the 2-month death rate shouldn't be used to predict the annual death rate, but I just annualised it to make the comparison, rather than converting annual death rate to monthly death rates (which I should have done).

So, accepting your correction, over a 2-month period, the "28+ survivors" cohort had a death rate of 5.5%.
Over the equivalent 2-month period, people aged 90+ have a death rate of 3.2%.
The point remains that people who are being classed as "recovered" or "didn't die of Covid" have experienced a death rate over 60% higher than that of 90+ year olds.

[servodude]

The point remains that people who are being classed as "recovered" or "didn't die of Covid" have experienced a death rate over 60% higher than that of 90+ year olds

If someone tested positive, was hospitalized two weeks later then put on a ventilator for two weeks a couple of days post admittance, before the plug was pulled and they finally succumbed... how would they be counted for the purposes of this set?

Recovered, because they lasted over 28days? and wrapped up in the 60day set?

-sd

[sg31]

Wrong thread

[sg31]

Wrong thread

[dealtn]

https://www.imperial.ac.uk/media/imperi ... 120820.pdf

quoted elsewhere but probably more relevant here

[johnhemming]

This study is based upon looking at individual cases
https://arxiv.org/pdf/2005.02090.pdf

Discussion
This paper does not prove that the peak in fatal infections in England and Wales preceded lockdown
by several days. Indeed the failure to undertake the sampling that could have gathered data to directly
measure infections early in the epidemic means that it will never be possible to be certain about timings,
given the substantial biases in clinical data other than deaths and fatal disease duration. What the results
show is that, in the absence of strong assumptions, the currently most reliable data strongly suggest
that the decline in infections in England and Wales began before full lockdown, and that community
infections, unlike deaths, were probably at a low level well before lockdown was eased. Furthermore,
such a scenario would be consistent with the infection profile in Sweden, which began its decline in fatal
infections shortly after the UK, but did so on the basis of measures well short of full lockdown.
These facts may have implications for the policies to be adopted in subsequent infection waves,
particularly given the peculiar ethical issues associated with lockdown. For example, a plausible estimate
of the life loss burden from an unmitigated COVID-19 epidemic in the UK is about 2 weeks per person5
.
A plausible lower bound on the UK life loss from the 2008 financial crisis and its aftermath is 7 weeks
per person6
. The economic shock from lockdown is substantially larger than 2008: Bank of England
projections suggest the largest shock for 100 or 300 years. Viewed another way, stringent suppression
measures might save 2 million UK life years, but the same UK population was on course to suffer around
200 million lost life years associated with economic deprivation and inequality before the COVID-19
crisis7
: carefully balanced policy is required to ensure that the suppression measures do not exacerbate
this by much more than one percent and lead to a net loss of life. Similarly the implied willingness to
pay to save a life year from COVID-19 appears to be an order of magnitude higher than the usual UK
National Institute for Health and Care Excellence threshold used for other diseases8
. Delayed health
interventions for serious conditions, although difficult to mitigate, represent a further life loss burden.

[Itsallaguess]

Not sure if a link to this interactive COVID-deaths map for England and Wales has been posted yet, but it can be used by postcode to see some interesting monthly data -

https://www.ons.gov.uk/visualisations/dvc957/map/index.html

Cheers,

Itsallaguess

[Itsallaguess]

The BBC have an article this morning showing the breakdown of August COVID cases in the UK broken down by age-group, and whilst there is clearly a large percentage of cases attributed to the younger end of the age-spectrum, I was surprised to see so many in the 30 to 60 age group as well -

The daily confirmed figures for coronavirus cases have shown a significant jump in the past two days. On Sunday they rose by nearly 3,000 and then on Monday by almost exactly the same again - the highest single day increases since May.

It means the UK is now seeing four times as many cases on average as it was in mid July. And with the impact of the return of schools still to come - not to mention the re-opening of universities next month - fears are growing that cases and then deaths could start spiralling.

The rise has been laid firmly at the door of young people. Around half of new cases in recent weeks have been diagnosed in people in their 20s and 30s.

Significant numbers of cases have also been identified in people in their 40s and 50s, as well as teenagers.



Source - https://www.bbc.co.uk/news/health-54064339

With the colder weather now well on it's way, and schools and colleges back up and running, it looks like we're going to have to get used to much higher levels of case-numbers, with a hope that those cases don't then require large-scale hospital interventions, and the recent swift upsurge in case numbers highlights the very fine line that policy-makers are having to run in terms of the balancing act between getting UK society moving again whilst still hoping for sensible behaviour whilst we do that.

Given that there seems to be an apparent willingness to allow case-numbers to rise in the UK without the reintroduction of more stringent UK-wide protocols, it's going to be difficult to firewall those 'allowable infections' within the age-group that are deemed to be OK, and so whilst the figures for COVID-related deaths are currently being maintained at a very low number, it looks like we might be in for a tough winter ahead for those much more susceptible sections of society if they don't look after themselves personally in a very stringent manner...

Cheers,

Itsallaguess

[tjh290633]

I feel that these figures might be better expressed as a proportion of the number of tests carried out.

That might confirm or dispel thougghts that the increased number of cases is a result of carrying out more tests.

Why are all these young people being tested? Do they have symptoms, or do they have to have a test for some other reason? A friend of mine had to have a test before he went into hospital for a procedure. Despite having a test site 10 to 15 miles from his home, he had to dig around before he got a spot allocated at the nearest one. When he got there it was virtually deserted, with people hanging around waiting for clients.

Why do they not use GP surgeries for the tests? Ours has a hot room at one of its 3 sites. I'm not sure what they do there. Probably answer the phone if they get a call. I suspect that it sees little if any use. I must ask the practice manager when I see her next.

TJH

[dspp]

deaths back and rising above trend

https://twitter.com/JKempEnergy/status/ ... 20/photo/2

regards, dspp

[dealtn]

It means the UK is now seeing four times as many cases on average as it was in mid July.

Do we know what the number of tests per day was in July, and compared to now?

[Itsallaguess]

It means the UK is now seeing four times as many cases on average as it was in mid July.

Do we know what the number of tests per day was in July, and compared to now?

I'm not sure, and there may well be an element of 'seek and ye shall find' in this increase in UK cases, but given that we've heard from Jonathan Van-Tam and Chris Whitty in the last 24-hours, both expressing serious concern over the rapid rise in UK cases in recent days, then I'd imagine they'd take that aspect into consideration before doing so, and perhaps caveat any concern with any testing-related influence, but they didn't seem to do that -

The latest "big change" in coronavirus infections across the UK is of "great concern", England's deputy chief medical officer has warned.

Prof Jonathan Van Tam said people have "relaxed too much" over the summer and "we have got to start taking this very seriously again".

If not, the UK will have "a bumpy ride over the next few months", he warned. He said that infections among younger people in EU countries had later filtered through to older age groups.

https://www.bbc.co.uk/news/uk-54065793

England's chief medical officer, Prof Chris Whitty, said Covid-19 rates are now rising, especially amongst people aged between the ages of 17 and 29.

He warned that if people stopped social distancing then "Covid comes back".

https://www.bbc.co.uk/news/uk-54066831

Cheers,

Itsallaguess

[dealtn]

I'm not sure, and there may well be an element of 'seek and ye shall find' in this increase in UK cases, but given that we've heard from Jonathan Van-Tam and Chris Whitty in the last 24-hours, both expressing serious concern over the rapid rise in UK cases in recent days, then I'd imagine they'd take that aspect into consideration before doing so, and perhaps caveat any concern with any testing-related influence, but they didn't seem to do that -

The latest "big change" in coronavirus infections across the UK is of "great concern", England's deputy chief medical officer has warned.

Prof Jonathan Van Tam said people have "relaxed too much" over the summer and "we have got to start taking this very seriously again".

If not, the UK will have "a bumpy ride over the next few months", he warned. He said that infections among younger people in EU countries had later filtered through to older age groups.

Well I guess it's their job, we would be extremely surprised if they weren't making such noises. I think the problem now is the authorities are short on bullets to deal with things as it is going to be extremely hard to re-introduce some of the harsher lockdown measures. If we do then there will no doubt be "a bumpy ride over the next few months" economically, with the severe consequences that will bring too.

Thankfully I am not in charge, but were I to be I would be putting the message out to my "health experts" to be ramping up the language.

[johnhemming]

Looking at hospital admissions there may be a smallish issue in the North West, but not otherwise.

[Mike4]

Looking on the bright side, as infections multiply and us older people start catching it and getting ill again, the statistical chances of any one person dying of it are greatly reduced due to better hospital treatment techniques.

On the dark side though, the long term sequelae will be as vicious as ever for the 17% of us who stay alive but don't get better from it.

[vrdiver]

It means the UK is now seeing four times as many cases on average as it was in mid July.

Do we know what the number of tests per day was in July, and compared to now?

If I remember correctly, wasn't July when we had ramped up to 100,000 tests per day? I saw a news comment yesterday saying we were now achieving 170,000 tests per day.

There was also the issue that the test stations could handle more tests, but the laboratories couldn't process them, hence the current reporting that some test stations are relatively quiet.

VRD

[Mike4]

If I remember correctly, wasn't July when we had ramped up to 100,000 tests per day? I saw a news comment yesterday saying we were now achieving 170,000 tests per day.

There was also the issue that the test stations could handle more tests, but the laboratories couldn't process them, hence the current reporting that some test stations are relatively quiet.

VRD

This whole thing doesn't stack up.

Firstly, the 100,000 tests per day claim by Mr Hancock was comprehensively debunked by the R4 programme "More or Less" which showed the true figure was closer to 70,000 per day IIRC. Counting home test kits posted out but not returned is simply frigging the figures for example, as well as other problems.

Secondly the press was full a few days ago of people being told to drive hundreds of miles to test centres to get a test as no appointments were available locally, yet on arrival finding fully staffed centres with nothing to do as no-one was turning up to be tested. Dido Harding has now walked away leaving this chaos for bigger and better things, heading up the new National Institute for Health Protection.

How does this square with claims of 170,000 tests being done per day? I bet Mr Hancock still uses weasel words like 'available tests per day' if one listens carefully. Hard to tell now they have cancelled the daily press briefings...

[servodude]

It means the UK is now seeing four times as many cases on average as it was in mid July.

Do we know what the number of tests per day was in July, and compared to now?

If I remember correctly, wasn't July when we had ramped up to 100,000 tests per day? I saw a news comment yesterday saying we were now achieving 170,000 tests per day.

There was also the issue that the test stations could handle more tests, but the laboratories couldn't process them, hence the current reporting that some test stations are relatively quiet.

VRD

If you're interested you can look at the Test and Trace stats site (for England) and find the period in question
- they consider both people and tests; because it's not uncommon to have multiple tests (because we know they're not 100% reliable)

The most recent is here (it lags real life by a week or two):
https://www.gov.uk/government/publications/nhs-test-and-trace-england-and-coronavirus-testing-uk-statistics-20-august-to-26-august-2020/weekly-statistics-for-nhs-test-and-trace-england-and-coronavirus-testing-uk-20-august-to-26-august#fn:1

Which states:
Between 20 August and 26 August, 452,679 people were newly tested for coronavirus (COVID-19), a decrease of 1% from the previous week. 6,732 new people had a positive result, an increase of 6% from the previous week1

Taking the two weeks in the middle of July

In week 8 of contact tracing (16 to 22 July):
366,397 people were newly tested for coronavirus (COVID-19) under pillar 1 (testing in hospitals and outbreak locations) and pillar 2 (national swab testing) in England
4,128 new people tested positive for COVID-19


and

Between 9 and 15 July 2020:
355,597 people were newly tested for coronavirus (COVID-19) under pillar 1 (testing in hospitals and outbreak locations) and pillar 2 (national swab testing) in England
3,953 new people tested positive for coronavirus (COVID-19)


so in that period people tested increased 27% for an increase in detected cases of 70%
- from the noises coming out of the CMO et al I suspect it looks to have got worse since

-sd