The maths of it all and what it means to Ireland

Hmmzis wrote: »

Our ICU teams must be doing something otherworldly, the survival rates are about 77% from ICU. That's a massive difference to other places where papers are being put in pre-print for achieving 60% survival rates.

https://www.medrxiv.org/content/10.1101/2020.05.03.20089318v1

Do we know what our teams are doing differently than anywhere else? If it's all down to care approach they should publish their protocols, it could save thousands of lives all over the worlde en before any special drugs are deployed.

idk. But Potentially we simply triage more / admit less lost cause cases?

AMKC wrote:

0.02806% That is if that's correct is the percentage so far of the population that the virus has killed in Ireland, I am quite sure that's right for that and for the amount of cases it is 0.4477 per cent or 0.45 per cent if you want to round it up.

The death rate is usually expressed as per 1000 or per 1M population.

Ireland is now at 0.291 per 1000 (1 in 3415 people).

This is a very poor result given the fundamentals - low population, low density, island on the periphery. Very badly managed unlike the uncritical media telling the people, who all believe "ah sure we're grand".

This is:
- About the same as the Netherlands (smaller area, the highest population density in Europe!)
- 2.7 times worse than Portugal (similar area, higher population)
- 3.3 times worse than Denmark (smaller area, similar population)
- 5 times worse than Austria (similar land area, double the population)
- 7.5 times worse than Norway (similar population, much larger area)
- 12 times worse than Czechia (similar land area, double the population)

UK should absolutely not be the bar for Ireland to gauge its results against, it's going to be the worst affected country in Europe, absolutely abysmal.

AMKC wrote:

The UK for cases was around 0.49 per cent when I done it last but could be higher or lower now.

UK death rate is 0.456 per 1000 but unlike most other countries this does not include care home and home deaths. The estimates done by FT and others using ONS data pointing at death rate of about 0.710 per 1000 which is brutal.

What it means for Ireland?

The government has failed to leverage all fundamental advantages of the country - island, low density, low population. Could have stopped and controlled all incoming traffic easily, check people and quarantine risk groups (see Taiwan - 26M people, only several deaths). Could have tested more, much more (see Korea). Could have mandated face masks (see vast majority of jurisdictions worldwide)- fatal mistake - still nothing on this. Instead chose very strict, long, highly economically damaging lockdown which didn't exactly work - see the other countries results above, their lockdowns were shorter and in some cases people were allowed to go to and from the work provided they wore a mask etc.

My prediction - if they do not up the game with massive testing and targeted quarantine and especially mandate masks for public spaces (all indoors and public transport as a minimum), we will see 20% unemployment, huge government deficit which we will all pay in taxes or austerity, cases and deaths will go up again in autumn and potentially if it happens along with a seasonal flu epidemic then the health care system may get overwhelmed, which would mean even more deaths. And there will be another lockdown, another deficit, another economic damage.

Do no take this lightly. Ireland would likely have 200k cases and 15k deaths if there was no shutdown. But repeated lockdowns are not an option.

Voltex wrote: »

Its a bit on the low side even with the age cohort and health profile of HCWs. But when you take into account other comparable cases i.e USS Theodore Roosevelt and the French carrier CdG we get a fairly consistent IFR of <<0.1%,

University of Bonn, Heinsberg study has an implied IFR of 0.37%. The lead author feels this a conservative number and may very well be closer to 0.2%

He also suggests we can take an IFR and extrapolate prevalence, which Ive been saying for a while. So if we take an IFR of 0.5% to allow for age and demographics, we could have/had >200k cases in Ireland, but we'll need to wait for seroprevalence studies to prove any such assumptions.

More and more studies are coming out that either test the entire population of a limited area, or else carry out random testing of a section of a population. Thats how you get an accurate picture of IFR. And they're showing a combined IFR of 0.28 on average:


https://docs.google.com/spreadsheets/d/1zC3kW1sMu0sjnT_vP1sh4zL0tF6fIHbA6fcG5RQdqSc/htmlview?pru=AAABchNlgG8*0znT152OBcjvTysrbg4e3w#gid=0

wakka12 wrote: »

Are lags in death rate taken into account in these studies? Or just the serum sample is compared with the number of deaths at that time?

For example, in the serum study in the Netherlands and Geneva, based on the serum prevalence then, the mortality rate today would be over 1% in both cases. Though perhaps infection rates also increased a lot since the date of the study too? I dont know

The Geneva study is quite interesting. Surveys over 8 weeks, starting April 6th showed prevalence:

Week 1 - 3.1%
Week 2 - 6.1%
Week 3 - 9.7%

Authors say theyll release more data over the coming weeks. Having said that, Geneva was hit fairly hard, so its probably to be expected they'd be at 20-30% now.

If we consider 15-18 days seroconversion lag, it would put infection period around first 1/3rd March ~ a week before peak.

I know the SC study got tonnes of blow-back from the stats people regarding specificity and selection bias. But from what I've read, there appears to be higher false negatives from the tests they used than false positives

Voltex wrote: »

So it appears there is serious and growing concern in the UK about the Imperial code used to model the epidemic.

https://lockdownsceptics.org/code-review-of-fergusons-model/

https://lockdownsceptics.org/second-analysis-of-fergusons-model/

I understand nothing about how these models work, but its enough to spook MP's, tech journos and programmers.

Then we have the real life manifestations.

1 - Nightingale hospitals stood down.
2- Deaths nowhere close to what was modelled.
3- No failure of hospital system.

Given NPHET uses the Imperial model (amongst others) in their projections, is this an area of concern?
https://twitter.com/SteveBakerHW/status/1259142970349105152

Deaths are far higher than what was recently predicted by the imperial model
https://theconversation.com/coronavirus-will-the-uk-really-have-highest-death-toll-in-europe-as-a-us-study-suggests-136017
The adjusted imperial model predicted 20,000 deaths in the UK if measures including school closures, stay-at-home orders and physical distancing for the entire UK population

Deaths are now in the region of 30,000 -40,000 and thousands more people will die
Hospitals were stretched to their limit in places like London during the peak, PPE shortages were huge concern. Certainly wasnt all rosy in the garden
The nightingale hospitals were a precautionary measure.

If youre referring to the early prediction in March of 250,000 deaths in the UK if no measures were introduced, well you certainly cant judge that figure as completely wrong so early into the pandemic.These predictions do not mean that we would have had 250,000 deaths at this exact point in time, i think many people forget. We are still only a couple of months in, it will up to two years more before a vaccine is found
If there are multiple waves in the UK of course deaths could rise up to 200,000 based on how many deaths UK is currently experiencing

MIT AI modelling seems to be bang on the money.

They have the US's mortality fairly accurate.

Interesting that they also estimate Ireland's total infected at 200k...just as I thought!!

https://covid19-projections.com/ireland

The model told us to do a thing or the results would be terrible.

So we did that thing.

And now the results arent terrible.

WHY DID THEY TELL US TO DO THAT THING?

DeVore wrote: »

That would presume that old people are more likely to catch it. They arent. They are more likely to *die* from it yes, but it spreads at the same rate through all ages as far as we know. Its just affects older people more.

Surely enclosed and populous environments like nursing and care homes (or prisons) are far more likely to catch it?

That means given the high proportion of older people in such environments, sadly they are more likely to catch it and have done.

DeVore wrote: »

That would presume that old people are more likely to catch it. They arent. They are more likely to *die* from it yes, but it spreads at the same rate through all ages as far as we know. Its just affects older people more.

Why wouldn't the elderly be more susceptible to getting infected when exposed?
They are more likely to have pre-existing conditions and weaker immune systems. But have less close social contact than younger people

yosemitesam1 wrote: »

Why wouldn't the elderly be more susceptible to getting infected when exposed?
They are more likely to have pre-existing conditions and weaker immune systems. But have less close social contact than younger people

Yes and for that reason they are more likely to die once infected. But as for actually contracting the infection in the first place, elderly people are not much more likely than other demographics to actually catch it

wakka12 wrote: »

If the virus was here a long time, the deaths would have consistently increased as infection rates rose. Elderly people are intermixed with the rest of society so there is no valid reason why this trend wouldnt occur

Are younger healthier people more likely to travel abroad, work, meet people?
If someone's sick with a bad cold, are they more likely to avoid elderly or at risk people but still show up at work?

I would think the answer is yes to both which makes it possible for a delay between initial circulation and hospitalisations/deaths.

niallo27 wrote: »

How do we know its accurate, they are miles off for Ireland today. Are these models recent or a few weeks old.

I don't know how these models work - all I can do is look at the results compared to what they predict, and so far the MIT model is fairly accurate (a lot more than the IHME model). I thinks its based on an SIR calculation and model, which is 1 of the 3 models used by Prof. Nolan

For a while now, I was taking an IFR of 0.5% and using that to guess our real infection rate. I felt we were +200K cases...which would also be consistent with seroprevalence surveys across Europe that indicate only 1/10 cases being picked up.

yosemitesam1 wrote: »

Are younger healthier people more likely to travel abroad, work, meet people?
If someone's sick with a bad cold, are they more likely to avoid elderly or at risk people but still show up at work?

I would think the answer is yes to both which makes it possible for a delay between initial circulation and hospitalisations/deaths.

I think this is correct to some degree. As I read more preprints on C19, I see how epidemiologists make the point that virus transmission is not homogeneous..its doesn't just get into the population moves through it evenly.

A paper that seems to be getting some attention was released by Oxford last week that suggests herd immunity may occur at rates far less than what traditional models suggest. They based this on the fact our social interactions are limited and predictable, therefore slowing the virus down after an initial period of rapid transmission.

Mike3287 wrote: »

Its accurate as we know from many countries now that death rate is about 0.50-1.00%

We have 1,500 or so deaths, worst we have 150,000 cases, best we have 300,000

Could be even higher amount of cases

Over 6500 health care workers have been infected, unfortunately 5 died

Death rate of 0.08%

If you think about it, herd immunity is very close in places like in UK, Sweden, Germany with amount deaths they have

If 50% of the population will have no symptoms, isnt that kind of 50% herd immunity?

Those countries are extremely far from herd immunity. The average person is not a health worker, the mortality rate is higher than 0.08% as you must consider the fact that a huge proprtion of our population is old and has health problems. These people are largely excluded from a sample of healthcare workers.