The maths of it all and what it means to Ireland

Hmmzis

yosemitesam1 wrote: »

False positives and false negatives are problems with the PCR test also.
The antibody testing shouldn't be about giving a definitive answer as to whether or not someone is immune but showing that the virus is nowhere near as dangerous as was originally suggested.
The initial surveys are showing 10-30+ times more spread than the official figures and false negatives are more likely than false positives.
So if they're scaled up further and continue to tell us that the virus mortality/hospitalisation rates are much lower than originally assumed, we could just let the virus run it's course in the general population

We would still need restrictions/countermeasures in place to avoid too fast a spread of it. As plenty of places have shown, it's capable of overrunning even the best healthcare systems in the world.

It should help shape those policies on restrictions better though.

Hmmzis

2u2me wrote: »

Can they also return 'false positive'?

If the specifity is crap then it could return a false positive for other similar proteins (antibodies against other coronaviruses, though that's a known issue and the labs and manufacturers should know how to isolate that).

yosemitesam1

Hmmzis wrote: »

If the specifity is crap then it could return a false positive for other similar proteins (antibodies against other coronaviruses, though that's a known issue and the labs and manufacturers should know how to isolate that).

The santa Clara county testing had less than 1% false positive and up to 30% false negatives

Hmmzis

yosemitesam1 wrote: »

The santa Clara county testing had less than 1% false positive and up to 30% false negatives

That's very similar to what the German case study was using, specificity >99%

normanoffside

DeVore wrote: »

Beware of antibody tests... they are quite inaccurate at times.

If you test 100k people and have a "false negative" rate of 95% accurate you are still giving 5k people the wrong result.

I'm working on a piece about this now but its complicated and hard to boiled down to easy communications

As far as I know they don’t give individual folks their results. They just use the stats to measure the general spread.

DeVore

DeVore wrote: »

Beware of antibody tests... they are quite inaccurate at times.

If you test 100k people and have a "false negative" rate of 95% accurate you are still giving 5k people the wrong result.

I'm working on a piece about this now but its complicated and hard to boiled down to easy communications

Thats not really how 95% accuracy works. A 5% false negative means that 5% of positives would be reported as negative. It the actual rate was 10% you would be giving 500 people the wrong result. Usually a test method has different false positive and a false negative rate, called alpha( false positive)and beta (false negative) error. When validating the test method you define your acceptable rate of each error type, and verify that the test method consistently achieves or exceeds the requirement

wakka12

https://news.sky.com/story/coronavirus-study-begins-to-track-covid-19-in-the-uk-population-11977336

UK to begin mass testing for antibodies within the community, 20,000 households will take part in the study.
The sample group will be tested every month for the next year

DeVore

2u2me wrote: »

Can they also return 'false positive'?

Yes, they can but if you test 100k people and the infection rate is 4% and you have a 95% accuracy then the number of false-positives is low because its 5% wrong of a much smaller number. Also, false positives have less impact (someone thinks they are sick when they arent).

normanoffside

DeVore wrote: »

Yes, they can but if you test 100k people and the infection rate is 4% and you have a 95% accuracy then the number of false-positives is low because its 5% wrong of a much smaller number. Also, false positives have less impact (someone thinks they are sick when they arent).

When they test 100k people in a random tests do you think they are actually giving back results to individuals?

Hmmzis

wakka12 wrote: »

https://news.sky.com/story/coronavirus-study-begins-to-track-covid-19-in-the-uk-population-11977336

UK to begin mass testing for antibodies within the community, 20,000 households will take part in the study.
The sample group will be tested every month for the next year

The 20,000 number is for throat swabs, that tests for an accute infection. The antibody tests will be done in 1000 households where no prior infection has been reported, no current symptoms are present and where nobody is shielding/cocooning.

mandrake04

DeVore wrote: »

Beware of antibody tests... they are quite inaccurate at times.

Serology Lab based Total Protein Assay is very accurate, they test serum using Electrochemiluminescence very sensitive.

You might be thinking of those finger prick Antibody tests that the UK and Spain had their pants pulled down with recently, they using full blood and trying get a reaction... IMO just laughable rubbish.

PCR tests extremely accurate, the issue there is mis-sample i.e the collector didn’t obtain good enough sample or was taken outside the sampling window.

If you swab too early you get a mis-sample as the virus possibly hadn’t time to settle in the nose, if you swab too late and the patient has recovered and the immune system possibly killed off the virus.

PCR Covid sample only has a recommended (by WHO) shelf life of something like 5 days if stored 2-8C. it can be frozen -70 but this might not happen until it’s near the end then you have to thaw out the other end etc. really you should be bunging them on the analysers within 24hrs.

IAMAMORON

wakka12 wrote: »

https://news.sky.com/story/coronavirus-study-begins-to-track-covid-19-in-the-uk-population-11977336

UK to begin mass testing for antibodies within the community, 20,000 households will take part in the study.
The sample group will be tested every month for the next year

It looks like blatant window dressing to me. They are going to randomly sample 20k out of 60M ( 3.3 ticks a month )for 12 months to set some sort of a benchmark or outlier for current virus population? All you can really hope to attain through such an exercise is a retort to an over inquisitive journalist at the weekly press conference?

Hypothetically there is an outbreak in say Walsall in September. 63 new clusters and 127 deaths in a week.

Journalist: " Mr minister for health, what have the gubberment done to protect the people of Walsall and what preventive measures are in place, has the government failed the people of Walsall this week?"

Gubberment Minister response " Well journalist as you are aware we announced last April a system of generating a national poll whereby we randomly tested 20,000 households to build a complex testing analysis mechanism to enable us to form a rapid response to any such outbreaks as they arise. We are delighted to say that once this outbreak has arisen we have been able to utilise the findings of these tests to implement a rapid response. We are saving lives. Blah blah blah lets all unite in sending our best wishes to the people of Walsall at this time, something something , virtue signal, thank you."

Case closed.

It is all window dressing and PR from now on in as far as I can see. Every other government in the world are at it.

normanoffside

So Most of the new cases in the last 2 days (about 66% from my calculations) are from nursing homes

https://twitter.com/FergalBowers/status/1253232672345817088

IAMAMORON

normanoffside wrote: »

So Most of the new cases in the last 2 days (about 66% from my calculations) are from nursing homes

https://twitter.com/FergalBowers/status/1253232672345817088

It is the new public relations Fire Fight. Expect a lot of media driven reactionary politics for the next month or so. Last month was testing, it is all nursing homes now.

87 % deaths are aged 70+ . 7% aged 60+ , 3.7% aged 50-60. 2% of people under the age of 50, who have tested positive, are dying. If you are under 50 with no underlying conditions you are safe to cross the road.

Is there any way of establishing the demographic of the actual tested population?

wakka12

'New York State Governor Cuomo said that preliminary findings from an antibody study conducted on 3,000 people at grocery stores across New York State found a 13.9% had coronavirus antibodies, suggesting a 13.9% actual infection rate statewide (21.2% in New York City), which translates to an estimate of about 2,700,000 actual cases in New York State (10 times more than the about 270,000 cases that have been detected and reported officially). Governor Cuomo acknowledged that the official count reported by New York State (which still is not including probable deaths as recommended by the new CDC guidelines) of about 15,500 deaths is "not accurate" as it doesn't account for stay at home deaths. Based on Worldometer's count (which includes probable deaths reported by New York City) of about 21,000 deaths and the 2,700,000 case estimate from the new antibody study, the actual case fatality rate in New York State could be at around 0.78%'

On worldometers

ShineOn7

Reddit daily digest



___________________________



New Cases: 936
Total Cases: 17607
New Death: 28
Total Deaths: 794 (3 deaths denotified)
Male: 421
Female: 373
Median Age: 83



MISSED THIS PART OF THE BRIEFING, CAN'T FIND NUMBERS?

• Deaths:
• Hospitalised + died in hospital:
• Died in ICU:
• Underlying conditions:
• Median age:
• Mean age: Range: 23-105

Of 332 in ICU

• Remain in hospital:142
• Discharged: 139
• Died: 50
• Underlying conditions: 277
• Median Age: 60

As of Tuesday 21st

• Cases: 16439
• In Hospital: 2424
• Total In ICU: 331
• Deaths: 789
• Clusters: 522
• Account for: 4096 cases
• Median age infected: 48
• Healthcare workers: 4545

Residential Settings

• All Residential Homes (Includes nursing homes): 319
• Nursing Homes: 191
• Total cases in residential homes: 2960
• Total cases in nursing homes: 2231
• Residential homes - deaths in hospitals: 433
• Nursing homes - deaths in hospital: 103

The below is based on information presented by Philip Nolan and the modelling
Parameters, and estimate of R0 as of this week

• Seen significant suppression in Ireland, monitored against all the different indicators
• Number of people who tested positive (5 day rolling average) is greatly reduced.
• External tests from Germany were backdated to date they occurred on. Many of those tests were backdated to late March, early April.
• Including these backdated tests, it shows that we have been detecting 500-600 cases in the community in a stable manner since the very strict public health measures were imposed
• The spread of the virus in the community stalled to almost nothing at one stage.
• Since the beginning of the week we can see a reduction in cases identified, highlighted by the downturn in the curve.
• In coming days we will see increase in cases due to the increased testing in residential care. Will see more cases than one might expect.
• However, there is a decline in general population, but increase in residential care
• Total people in hospital - peaked at 900 last week, but it is now decreasing. More being let home than people admitted
• ICU grew steadily March, peaked at roughly 160, was steady in early April, and now declining
• Hospital admissions peaked at 90 patients a day

Disease has been stabilised or suppressed

• Looking very closely at residential cases.
• Looking very carefully at when the disease was actually detected. Swab would have been taken roughly at the time when the disease presented.
• Nursing home testing shows that it took the disease some time to get into nursing homes. Almost 2 weeks after its rapid rise in the general population, it rapidly grew in nursing homes. It is continuing to rise in nursing homes.
• Disease emerged in general population before residential care.
• Disease is suppressed in general population
• Not clear that it is suppressed in residential home
• Recent data shows that it is declining in residential homes
• Driving the disease out of the general public is the best protection for the residential home patients

• Number of deaths a day have been graphed by the date that they occurred
• Number of deaths per day in population and long term residential homes has been relatively stable since the end of the first week in April, a final indication that the disease is under control
• Growth rates in confirmed cases - in early stages of epidemic the disease was growing at 33% per day. This has declined in the course of March and early April. Growth rate since early April has been pretty much zero (5 day average). This is creeping up due to the extra testing being carried out in long term residential homes.

• Can estimate/confirm the number of people who have recovered from the disease, based on : (# released from hospital) + (Those who had a positive test 14 days ago and haven't been hospitalised/died).
• 2/3rds of people diagnosed with Covid have recovered
• One of the key metrics to know before we consider changes in the restrictions - disease needs to be managed to a very low level (we are close to this, but this will take some time to reduce hospitalised/icu/community level to a low number)
• We need to keep the R0 below 1, we can do this, and we can force this down as needed

• R0 based on 4 estimates. All 4 estimates confirm with a high level of confidence that the R0 is between 0.5 and 0.8
• Imperial College London model - tell the number of cases, the number of deaths, and the point at which you introduced measures. This model showed before restrictions, R0 = 2-4 Since strong social distancing measures, R0 <1, 0.54 (95% confident)
• Half population is transmitting it to no one, other half are transmitting it to only one other person
• R0 = 0.8 - 600 cases a day - our data follows the model well (ignored data from last two days, as there is a difference in the testing regime - detecting cases that would not have been tested before)
• If on May 5th restrictions are relaxed and people still follow social distancing, the numbers will rise slowly. Reintroducing strict measures will reduce numbers again
• If R0 = 1.6, people are more relaxed, contact each other more. Will show a peak much higher than what we have seen. Strict reintroduction will control the diseased
• R0 =2.4, then we cannot cope.
• These model projections show us that we need to be exceptionally cautious so that we don't have something that significantly increases the transmission of the virus

Questions

• Do you feel like you are losing the public lately? Have the majority of the public with us. We have evidence to show that we are having some extra movement. Modelling shows that if we do very little, we will see large growth of the virus. If reproductive number climbed up to high 1, beyond 2, we would have significant impact on health of public, and on the HSE
• We have a long way to go still, especially with the increase in cases that were picked up today due to the residential homes testing To contemplate lifting the restrictions, we need to make sure that the community residential homes don't infect the general community We need to break the chain of transmission in the nursing homes in the way that we have done so in the general community.
• 4 out of 5 deaths have occurred in the East. Can we ease the measures in the rest of the country? They have looked closely at the question earlier in the disease, but ultimately decided the right thing to do was to implement measures across the entire country. This would help prevent the spread of disease in other parts of the country. Cannot expect restrictions will become regionalised.
• For each case detected, how many go undetected? Worldwide thought that for every case detected, 1 goes undetected. Serological studies have been prepared, once we have a good reliable test. Will random select people to do a blood sample to confirm the prevalence of the disease in the community.

Jurgen Klopp

3,830 of Irish healthcare workers have been confirmed infected

159 required hospital that's 4.1% of confirmed cases

24 ICU 0.63%

5 deaths 0.13%

You'd assume a nice mix of ages there too and I'm sure some with various conditions not much different to the general population?

ShineOn7

From Reddit - interesting bit I noticed and highlighted in red



______________________


New Cases: 577 - 245 in respect of nursing homes
Total Cases: 18184
New Death: 37
Total Deaths: 829 * 2 denotifications
Male: 441
Female: 388
Deaths

• Hospitalised + died in hospital: 406
• Died in ICU: 53
• Underlying conditions: 715
• Median age: 83
• Mean age: 81
• Range: 23-105

Of 340 in ICU

• Remain in hospital: 142
• Discharged: 144
• Died: 53
• Underlying conditions: 283
• Median Age: 30

As of Weds 22nd

• Cases: 17420
• In Hospital: 2486
• Total In ICU: 338
• Deaths: 821
• Median age infected: 49
• Healthcare workers: 27%

Update on Healthcare Workers - Data up to Saturday 18th April

• 3830 cases *159 hospitalised
• 24 admitted to ICU
  • 5 deaths

Residential Settings

• All Residential Homes (Includes nursing homes): 336 clusters
• Nursing Homes: 198 clusters
• Total cases in residential homes: 3277 - 305 hospitalised
• Total cases in nursing homes: 2500 - 208 hopsitalised
• Residential homes - deaths in hospitals: 452 lab confirmed - 124 probable cases
• Nursing homes - deaths in hospital: 376 lab confirmed - 112 probable
• 576 total across all community residential homes, when including confirmed and probably
• 488 across all nursing homes, when including confirmed probable
• 109 died in hospital
• 91 out of 376 nursing home deaths occurred in hospital

Questions

• ECDC has asked that all countries report the probable and confirmed deaths together. Ireland has been doing this already. We are one of 4 countries doing this.
• Probable deaths what does this mean? A death of someone who didn't have a swab taken, but the doctor believes that the person was likely to have had Covid.
• Decision in principle has been decided on case definition. What that case definition will be is in line with the ECDC advice: Fever, recent onset of cough, shortness of breath. An algorithm will be used to try and see how many numbers that will be. There will still be prioritisation of certain groups (elderly, healthcare workers) even when the criteria is expanded.

Blut2

Jurgen Klopp wrote: »

3,830 of Irish healthcare workers have been confirmed infected

159 required hospital that's 4.1% of confirmed cases

24 ICU 0.63%

5 deaths 0.13%

You'd assume a nice mix of ages there too and I'm sure some with various conditions not much different to the general population?

Theres a lot of evidence that the more exposed you are to the virus (ie in either time duration, or else from being around multiple people with the virus) the more severe you get it yourself - due to being exposed to a heavier viral load.

So given that, if anything, the heathcare workers are likely having worse outcomes than the population in general will. I'd imagine thats whats responsible for the comparatively high hospitalization rate.

Still good news overall with the very low death rate, though.

ShineOn7

Blut2 wrote: »

Theres a lot of evidence that the more exposed you are to the virus (ie in either time duration, or else from being around multiple people with the virus) the more severe you get it yourself - due to being exposed to a heavier viral load.

I've been reading everything on this topic and that's a first for me



Have you any links from credible websites?

Blut2

ShineOn7 wrote: »

I've been reading everything on this topic and that's a first for me

Have you any links from credible websites?

https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(20)30232-2/fulltext

Or if you want something a bit more lay accessible:

https://www.nytimes.com/2020/04/01/opinion/coronavirus-viral-dose.html

ShineOn7

Could someone caulculate the mortality rate for


(all without underlying conditions)


In Ireland, the mortality rate for:

- Under 65s
- Under 70s
- Those aged 70 to 80
- Those aged over 80


Worldwide, the mortality rate for:

- Under 65s
- Under 70s
- Those aged 70 to 80
- Those aged over 80

Blut2

ShineOn7 wrote: »

Could someone caulculate the mortality rate for

(all without underlying conditions)

In Ireland, the mortality rate for:

- Under 65s
- Under 70s
- Those aged 70 to 80
- Those aged over 80

Worldwide, the mortality rate for:

- Under 65s
- Under 70s
- Those aged 70 to 80
- Those aged over 80

I think you may struggle to get such specific data - I've not seen it anywhere yet, anyway.

This might be of interest along similar-ish lines though, its a great report on deaths in Italy to date:

https://www.epicentro.iss.it/en/coronavirus/bollettino/Report-COVID-2019_23_april_2020.pdf

Key findings include:

- only 3.6% of all deaths, across all age ranges, had no serious underlying health condition (diabetes, cancer etc). 96.4% of deaths had a serious underlying health condition.
- 60% of people who died had 3 or more(!) underlying serious health conditions
- Median age of death of 81
- only 9 deaths total of people under 30 in Italy from coronavirus to date

Its just not killing young people at all, and even amongst old people its very rarely killing people with no underlying health conditions.

ShineOn7

Blut2 wrote: »

I think you may struggle to get such specific data - I've not seen it anywhere yet, anyway.

This might be of interest along similar-ish lines though, its a great report on deaths in Italy to date:

https://www.epicentro.iss.it/en/coronavirus/bollettino/Report-COVID-2019_23_april_2020.pdf

Key findings include:

- only 3.6% of all deaths, across all age ranges, had no serious underlying health condition (diabetes, cancer etc). 96.4% of deaths had a serious underlying health condition.
- 60% of people who died had 3 or more(!) underlying serious health conditions
- Median age of death of 81
- only 9 deaths total of people under 30 in Italy from coronavirus to date

Its just not killing young people at all, and even amongst old people its very rarely killing people with no underlying health conditions.

This information feels very outdated. Several people in their 40s and younger with no underlying conditions have died from this

It killed a 13 year old in Britain who had no underlying conditions

wakka12

I think that stats dont lie, but media focus dispoportionally on the deaths of young and healthy people. Having said that the emergence of recent stories in the US of a correlation between several young COVID patients with few or no symptoms suffering strokes is extremely worrying

ShineOn7

wakka12 wrote: »

I think that stats dont lie, but media focus dispoportionally on the deaths of young and healthy people. Having said that the emergence of recent stories in the US of a correlation between several young COVID patients with few or no symptoms suffering strokes is extremely worrying

Jesus

Hmmzis

wakka12 wrote: »

I think that stats dont lie, but media focus dispoportionally on the deaths of young and healthy people. Having said that the emergence of recent stories in the US of a correlation between several young COVID patients with few or no symptoms suffering strokes is extremely worrying

Need to dig around, but there was a concern regarding the obesity in the general US population being a serious risk factor. The Italians noticed these complications as well, but from my understanding in much fewer cases.

Blut2

ShineOn7 wrote: »

This information feels very outdated. Several people in their 40s and younger with no underlying conditions have died from this

It killed a 13 year old in Britain who had no underlying conditions

Those stats are up to date as of April 23rd - thats when the linked report was published. They're extremely up to date.

What weight/size was the 13 year old? As Hmmzis mentions, a lot of the "no underlying conditions" young people who've died have actually been obese - which is a key factor in chance of death. That or they have underlying health conditions that they/their family were unaware of at the time of death.

And as wakka12 mentions the media does love to focus on any young deaths for the shock factor, so the odd anecdotal one here or there doesn't really prove anything - the verified statistics, based on tens of thousands of deaths, don't lie. Otherwise healthy young people are more likely to die from choking on their food at dinner tonight than they are from corona.

PommieBast

Looking over some of the graphs it seems like Europe is following what looks more like a Ricean distribution than the Gaussian assumed in models and seen in Korea. Means the tail-off will not be as steep as hoped

ShineOn7

PommieBast wrote: »

Looking over some of the graphs it seems like Europe is following what looks more like a Ricean distribution than the Gaussian assumed in models and seen in Korea. Means the tail-off will not be as steep as hoped

In layman terms could you describe the difference between the two?