The maths of it all and what it means to Ireland

bb1234567

Seweryn wrote: »

For Cov.

Could you back up your comment about global death toll being underestimated seweryn. it's really so exhausting to hear such dismissive comments daily that go completely unchallenged despite providing little in the way of any meaningfil contribution to discussion. Your comment applies to global death toll, yes it may be overstated in Ireland and Belgium and UK, but that is a very small region of the globe overall. There are many other countries throughout much of the rest of the world, majority of the developing world, where there is very strong evidence that the death toll is very substantially underestimated. So could you please provide some information back up your statement, if not comments such as yours are just distracting and unhelpful in the discussion.

Andrewf20

bb1234567 wrote: »

Well I mean since mid August Spain added almost 4,000 deaths but okay..obviously nowhere near what it was in April but it's still a significant increase in excess mortality in Spain

4000 sounds a lot alright. I did some maths which I hope is right. 4000 divided by 50 days (approx 7 weeks since mid August)is an average of 80 deaths per day. Population is 47 million. 47 million divided by 80 is on average 1 extra death in 587,000 people per day. If that's correct, it still doesn't feel like a huge increase tbh even with rates rising.

ShineOn7

Blut2 wrote: »

Corona is currently a distant 4th most deadly flu of the last century, nevermind "once in a century". Its really not that special.

I'll let the 1,800 Irish families who lost loved ones to Covid know

Thanks

ShineOn7

Laura on Reddit Ireland transcribing today's data

Modelling and Statistics

5 Day Average

• Last Thursday: 370 cases per day, on average
• Today: 506 cases per day, on average

Positivity Rate

• Up to last Thursday: 3%
• Up to today: 4%

Deaths (based on date of death)

• August: 5
• September: 34
• October so far: 8

All key indicators of disease severity are increasing

• Hospitalisations
• ICU
• Deaths

• Incidence rate is of concern in every county
• Proportion of case amongst younger people continues to grow
• Cases within the 19-24 is very worrying worrying, high incidence
• Cases in 65+:
  • Sept: 10-15/100,00
  • Today: 30/100,000

Reproduction number

• R=1.2
• Likely Dublin is closer to 1
• Rest of country 1.5
• Disease spreading very rapidly, particularly outside Dublin

Modelling

• Model has proven an accurate prediction
• Re-run the model yesterday
• At current R=1.2-1.4 - We can expect up to 1100-1500 per day by end of first week in Nov
• Growth rate of the country as whole 4% - double every 18 days
• Recent variations in Dublin make growth rate more difficult to determine.
• Growth rate outside of Dublin much higher than 4%

Andrewf20

ShineOn7 wrote: »

Laura on Reddit Ireland transcribing today's data

The numbers are expected to increase as we head into winter like in any year. 8 deaths in October sounds quite low still. Cases rising since July.

Just found an article to gain some perspective, from 2015:

https://www.irishtimes.com/life-and-style/health-family/medical-matters-cold-weather-and-fuel-poverty-cause-2-800-excess-deaths-1.2099569

“As the island of Ireland currently has the highest levels of excess winter mortality in Europe, with an estimated 2,800 excess deaths during each winter..."

Im not sure what winter is defined as exactly, assuming its the standard 3 months, then 2800 / 3 months = 933 average excess deaths per month is considered normal.

Another article here:

https://www.irishnews.com/news/northernirelandnews/2018/12/20/news/excess-winter-deaths-hit-highest-ever-total-1512563/#:~:text=There%20were%206%2C404%20deaths%20recorded,640%20the%20previous%20two%20winters.&text=Of%20the%201%2C500%20excess%20deaths,occurred%20in%20the%20over%2D85s.

6400 excess deaths in the north in winter 2017/18. Unusually high in fairness, but no notion of lockdowns.

bb1234567

Andrewf20 wrote: »

The numbers are expected to increase as we head into winter like in any year. 8 deaths in October sounds quite low still.

Just found an article to gain some perspective, from 2015:

https://www.irishtimes.com/life-and-style/health-family/medical-matters-cold-weather-and-fuel-poverty-cause-2-800-excess-deaths-1.2099569

“As the island of Ireland currently has the highest levels of excess winter mortality in Europe, with an estimated 2,800 excess deaths during each winter..."

Im not sure what winter is defined as exactly - if its 3 months, then 2800 / 3 months = 933 average excess deaths per month is considered normal.

That is a surprisingly high number but important to realise it's 2800 deaths on the island of Ireland. No idea why NI and rep numbers were lumped together but they explicitely said on 'the island of Ireland' so obviously they are added together in this instance.

So it's about 2100 in the rep of Ire. Still really high mind you and a lot hgiher than I thought would be our usual winter excess

bikeman1

Does anybody know how they reflect the numbers when there are late reporting or subsequent de-notifications of deaths?

Are the numbers for each month reclassified to give the true figure or are they just going on the headline numbers.

For example, referenced above there were 5 deaths in August. However on 3rd October we were told of 10 deaths with 8 “prior to September 2020”. Is there any detail on when these deaths actually occurred. It is conceivable to me that some were in August. Do the 5 deaths quoted for August include these late reported deaths?

Similarly so far in October the following numbers have been announced / de-notified:

1/10 4 deaths / 2 de notified
2/10 1 death / 6 de notified
3/10 2 deaths (plus 8 prior to Sep) / 1 de notified
4/10 No change
5/10 No Change
6/10 1 death
7/10 5 deaths

The reason I ask is statistics can take a whole different meaning depending on where you put the numbers. Coming off a “low base” in August can make September look way worse than it was. Where do the de notifications fit into it?

Also, why are there so many de-notifications. Surely a positive swab has to be there BEFORE confirming that the person died with Covid? It is well known that they have no problem waiting for results to come back in many cases.

IAMAMORON

bikeman1 wrote: »

Does anybody know how they reflect the numbers when there are late reporting or subsequent de-notifications of deaths?

Are the numbers for each month reclassified to give the true figure or are they just going on the headline numbers.

For example, referenced above there were 5 deaths in August. However on 3rd October we were told of 10 deaths with 8 “prior to September 2020”. Is there any detail on when these deaths actually occurred. It is conceivable to me that some were in August. Do the 5 deaths quoted for August include these late reported deaths?

Similarly so far in October the following numbers have been announced / de-notified:

1/10 4 deaths / 2 de notified
2/10 1 death / 6 de notified
3/10 2 deaths (plus 8 prior to Sep) / 1 de notified
4/10 No change
5/10 No Change
6/10 1 death
7/10 5 deaths

The reason I ask is statistics can take a whole different meaning depending on where you put the numbers. Coming off a “low base” in August can make September look way worse than it was. Where do the de notifications fit into it?

Also, why are there so many de-notifications. Surely a positive swab has to be there BEFORE confirming that the person died with Covid? It is well known that they have no problem waiting for results to come back in many cases.

I think that you are really taking it too far here.

At this point shifting goalposts over August or September is immaterial. There are countless ways the information gets messed around. The golden rule is that the HSE is a shambles and all things considered, giving the nature of having to report everyday to deadlines and the pressure they are under, there will be loads of phuck ups. But 3-4 deaths either side of 2 months will not be substantially blurring up the standard deviation?

bikeman1

IAMAMORON wrote: »

I think that you are really taking it too far here.

At this point shifting goalposts over August or September is immaterial. There are countless ways the information gets messed around. The golden rule is that the HSE is a shambles and all things considered, giving the nature of having to report everyday to deadlines and the pressure they are under, there will be loads of phuck ups. But 3-4 deaths either side of 2 months will not be substantially blurring up the standard deviation?

The NPHET are using the deteriorating numbers as reason behind their advice to go straight to level 5. The commentary around that all refers to Sep and Aug figures and where we have come from and to. Deaths are the key metric in all of this in my view.

Announcing deaths and then saying a few weeks later, ah sorry lads they were just suspected at the time, turns out they weren’t. Oh yeah you are all ok level 5 and 300,000 out of work! It’s serious stuff, people’s mental health and livelihoods depend on advice.

bb1234567

bikeman1 wrote: »

The NPHET are using the deteriorating numbers as reason behind their advice to go straight to level 5. The commentary around that all refers to Sep and Aug figures and where we have come from and to. Deaths are the key metric in all of this in my view.

Announcing deaths and then saying a few weeks later, ah sorry lads they were just suspected at the time, turns out they weren’t. Oh yeah you are all ok level 5 and 300,000 out of work! It’s serious stuff, people’s mental health and livelihoods depend on advice.

I think from the death toll in March/April remaining relatively unchanged with only minor reduction through denotification 5-6 months later shows that the vast vast majority of deaths which are suspected are suspected for good reaosn and do turn out to be covid. There's not enough alterations made to prove in anyway that the daily death tolls are misleading them as their guidelines

I would also strongly disagree that deaths are the determinig metric anyway, rising hospitalisations but especially ICU admissions are quite clearly the determining factor in restrictions. Deaths have still not really risen, or anywhere near enough to justify what's being talked about, hospitalisations are rising at a concerning rate though

ShineOn7

For those for a real eye for this, the CDC have a page for Mathematical Modeling https://www.cdc.gov/coronavirus/2019-ncov/covid-data/mathematical-modeling.html

patnor1011

bb1234567 wrote: »

Well I mean since mid August Spain added almost 4,000 deaths but okay..obviously nowhere near what it was in April but it's still a significant increase in excess mortality in Spain

35k people die in spain every month - average from last year. Some of those 4k or most of them would most likely die from flu or some other complication anyway.

IAMAMORON

I cannot help but think that the increase in testing and the kids back in school has led to the current spike.

Where can I get figures on testing, is there any available?

bb1234567

IAMAMORON wrote: »

I cannot help but think that the increase in testing and the kids back in school has led to the current spike.

Where can I get figures on testing, is there any available?

Hospitalisations are increasing, if it was down to testing they wouldnt be spiking somultaneously wiyh cases

McGiver

OK, the narrative is "the numbers are rising" but I completely miss a decent analysis of those numbers. The folk is just running in panic repeating "the numbers are rising!"

What is missing in the narrative:
- April peak was 250 cases/100k
- Current number is less than 50% of that at 110 cases/100k
- Death rate is not increasing proportionally with cases/capita as it did in April (the peak was 150/100k), in fact it's almost not rising at all (it's at 3.3/100k)
- Positivity rate/100k is at % (2.3%) and also not rising proportionately as it did in April (the peak was 27%), in fact it's not rising significantly at all
- Hospital admissions are at 0.2/100k vs April peak of 6.7/100k, not rising at all so far
- Different age group is affected - it is more 15-50 now
- All of the above bar positivity is seen in most other EU countries (bar Greece, Hungary, Czechia, Croatia, Slovakia)
- The numbers are increasing in almost all EU countries with the sole exception of Sweden as more testing is done/schools open/borders opened

And for context:
- Belgium, France, the Netherlands are at 250+ cases/100k, Czechia is 350/100k, Spain at 300/100k and their positivity is increasing as well, while death rate isn't increasing bar Czechia (because it was extremely low before) - so the numbers seem much worse than Ireland now, yet no panic anywhere really

Where are we with hospitalisations, and especially compared with April/May numbers? EDIT - found the numbers

ShineOn7

bb1234567 wrote: »

Hospitalisations are increasing, if it was down to testing they wouldnt be spiking somultaneously wiyh cases

Are they @ 8% of the number of cases like in April?

Genuinely curious

patnor1011

McGiver wrote: »

OK, the narrative is "the numbers are rising" but I completely miss a decent analysis of those numbers. The folk is just running in panic repeating "the numbers are rising!"

What is missing in the narrative:
- April peak was 250 cases/100k
- Current number is less than 50% of that at 110 cases/100k
- Death rate is not increasing proportionally with cases/capita as it did in April (the peak was 150/100k), in fact it's almost not rising at all (it's at 3.3/100k)
- Positivity rate/100k is at % (2.3%) and also not rising proportionately as it did in April (the peak was 27%), in fact it's not rising significantly at all
- Hospital admissions are at 0.2/100k vs April peak of 6.7/100k, not rising at all so far
- Different age group is affected - it is more 15-50 now
- All of the above bar positivity is seen in most other EU countries (bar Greece, Hungary, Czechia, Croatia, Slovakia)
- The numbers are increasing in almost all EU countries with the sole exception of Sweden as more testing is done/schools open/borders opened

And for context:
- Belgium, France, the Netherlands are at 250+ cases/100k, Czechia is 350/100k, Spain at 300/100k and their positivity is increasing as well, while death rate isn't increasing bar Czechia (because it was extremely low before) - so the numbers seem much worse than Ireland now, yet no panic anywhere really

Where are we with hospitalisations, and especially compared with April/May numbers? EDIT - found the numbers

This virus cant be stopped with lockdowns. Any lockdown just delay inevitable. It also speeds up economic demise with all which comes with it.

Wanderer78

patnor1011 wrote:

This virus cant be stopped with lockdowns. Any lockdown just delay inevitable. It also speeds up economic demise with all which comes with it.

Governments have the capabilities to soften the negative economic impacts significantly, it can be done, but will they.....

CIARAN_BOYLE

ShineOn7 wrote: »

Are they @ 8% of the number of cases like in April?

Genuinely curious

No. The hospitalisation rate in March was distorted by hospitals being used as isolation facilities for those who needed them.

We have 159 in hospital atm. I've no idea about discharges but the amount in hospital is growing every day atm.

ShineOn7

John Campbell says in a video this week that the IFR worldwide is now at 0.3%

It was at 0.5% just 6-8 weeks ago and I'll take any bit of good news going this week

So IFR @ 0.3% = 3 times more now more lethal than the Flu (IFR @ 0.1%), when back in May is was thought to be upto 14 times more lethal (source: https://www.healthline.com/health-news/why-covid-19-isnt-the-flu)

It's still important to remember that, last I checked, Covid is vastly more contagious than the Flu. There's also Long Covid, which seems to be reported more and more lately

https://www.youtube.com/watch?v=oDlNdit0hKA

McGiver

ShineOn7 wrote:

So IFR @ 0.3% = 3 times more now more lethal than the Flu (IFR @ 0.1%), when back in May is was thought to be upto 14 times more lethal (source:

Flu IFR is 0.04 if you add asymptomatic estimated...

CalamariFritti

But the WHO thinks we have 750m infected. And 1m dead. So three times the flu seems not off by magnitudes anyway.

ShineOn7

McGiver wrote: »

Flu IFR is 0.04 if you add asymptomatic estimated...

We can't really base calculations on this though. Asymptomatic stats on Covid have had varying wild swings in guesses since April

I found another article here https://theconversation.com/now-everyones-a-statistician-heres-what-armchair-covid-experts-are-getting-wrong-144494

First, if we compare the typical flu IFR of 0.1% with the most optimistic COVID-19 estimate of 0.25%, then COVID-19 remains more than twice as deadly as the flu.

But there's also this and it stresses how much more infectious Covid is versus the Flu

Flu’s R₀ is about 1.3. Although COVID-19 estimates vary, its R₀ sits around a median of 2.8. Because of the way infections grow exponentially (see below), the jump from 1.3 to 2.8 means COVID-19 is vastly more infectious than flu.

ShineOn7

CalamariFritti wrote: »

But the WHO thinks we have 750m infected. And 1m dead. So three times the flu seems not off by magnitudes anyway.

I don't put much faith in WHO anymore

They're just a political chess piece of China's at this stage

2ndcoming

Sure they are :rolleyes:

Blut2

https://twitter.com/oconnellhugh/status/1317582515192975360?s=21


Some topical not great maths of it all for Ireland this week.

ShineOn7

Blut2 wrote: »

https://twitter.com/oconnellhugh/status/1317582515192975360?s=21


Some topical not great maths of it all for Ireland this week.

Not what 99% of this thread has been about at all

There's dozens of threads you'll find more people to bite

ShineOn7

Just muted someone on social media for comparing the stats of automobile accidents to Covid statistics

They weren't someone I'd have deemed as stupid until today either

ShineOn7

Can anyone with an analytical eye give this a look?

On page 8 of this PDF it gives the % amount of those hospitalized in Ireland with Covid in different age brackets

https://www.hpsc.ie/a-z/respiratory/coronavirus/novelcoronavirus/casesinireland/epidemiologyofcovid-19inireland/COVID-19_Daily_epidemiology_report_(NPHET)_20201027%20-%20website.pdf

But on the next page it gives completely different % numbers

Example:

Page 8 says of those aged 35-44, 7.99% have been hospitalized. Which sounds high for that age bracket

But on the next page, for the same age bracket, it says 3.69% have been hospitalized. This one seems more realistic

Or am I just reading page 8 wrong and I need more tea?

ShineOn7

ShineOn7 wrote: »

Can anyone with an analytical eye give this a look?

On page 8 of this PDF it gives the % amount of those hospitalized in Ireland with Covid in different age brackets

https://www.hpsc.ie/a-z/respiratory/coronavirus/novelcoronavirus/casesinireland/epidemiologyofcovid-19inireland/COVID-19_Daily_epidemiology_report_(NPHET)_20201027%20-%20website.pdf

But on the next page it gives completely different % numbers

Example:

Page 8 says of those aged 35-44, 7.99% have been hospitalized. Which sounds high for that age bracket

But on the next page, for the same age bracket, it says 3.69% have been hospitalized. This one seems more realistic

Or am I just reading page 8 wrong and I need more tea?

On page 8 the7.99% is the percentage of total hospitalisations that were in 35-44 year olds, on the next page it’s % of 35 to 44 year olds who have been hospitalised