Scrutable

Woodchopper wrote: ↑

Tue Apr 07, 2020 9:18 pm

raven wrote: ↑

Tue Apr 07, 2020 8:53 pm

shpalman wrote: ↑

Tue Apr 07, 2020 7:56 pm

It's here: https://covid19.healthdata.org/united-kingdom

799 feels way to low for the total number of ICU beds in a country anyway though so I'm confused about that number.

It's possibly unoccupied ICU beds. We have something like 4000, and at the standard 80% occupancy that leaves about 800 free to use.

That’s what I’m assuming. But if so it ignores the capacity that’s been built over the past few weeks.

The number of intensive-care beds has been doubled to nearly 10,000

"More than 2,000 of them are still available."

Italy peaked at around 4,000 patients in intensive care across the country, and this number has been falling for a few days.

So two weeks after Italy finished its field hospital in an Expo, which it ended up not needing, the same thing is happening in the UK.

Still, better to have them and not need them.

https://covid19.healthdata.org/united-kingdom seems to have made its prediction a bit more reasonable even if it still thinks there aren't enough beds.

Maybe. We only have "enough" capacity because we have cancelled huge swathes of non-Covid medicine, allowing people with other conditions to die. Once we are confident that Covid rates aren't increasing, we can rebalance that and find that there is still demand for far more than the baseline number of beds. Although to the extent that the limiting factor is the trained staff hours, maybe the extra beds themselves won't be that useful.

Yes. Need to return the main NHS to cancer, heart, diabetes, hip replacements, etc - millions are suffering and some are dying due to avoidance of healthcare.

Which means much much more training of basic nurses to be Covid specific carers.

lpm wrote: ↑

Wed Apr 15, 2020 9:30 pm

Yes. Need to return the main NHS to cancer, heart, diabetes, hip replacements, etc - millions are suffering and some are dying due to avoidance of healthcare.

Which means much much more training of basic nurses to be Covid specific carers.

And underpinning it all, we have to get our act together on rapid covid testing. With this sorted, it becomes much easier to keep the cancer/diabetes/etc patients safe while treating them.

Supporting this could potentially be the longer term value of the Nightingale hospitals, as a way to segregate off the covid-infected. (Papworth Hospital, now a specialist heart/lung centre, started off as a TB sanatorium.)

I'm sure I once knew the answer to this, but I've forgotten. How come there's such a shortage of tests in Europe and the USA, when places like South Korea managed to get hold of shitloads at short notice?

Bird on a Fire wrote: ↑

Wed Apr 15, 2020 11:31 pm

I'm sure I once knew the answer to this, but I've forgotten. How come there's such a shortage of tests in Europe and the USA, when places like South Korea managed to get hold of shitloads at short notice?

https://www.cnn.com/2020/03/12/asia/cor ... index.html

Herainestold wrote: ↑

Thu Apr 16, 2020 2:32 am

Bird on a Fire wrote: ↑

Wed Apr 15, 2020 11:31 pm

I'm sure I once knew the answer to this, but I've forgotten. How come there's such a shortage of tests in Europe and the USA, when places like Southu Korea managed to get hold of shitloads at short notice?

https://www.cnn.com/2020/03/12/asia/cor ... index.html

Thanks for this article. It's an interesting story that I hadn't heard. But I think it only answers half the question.

Why aren't countries like the UK licensing the Korean tests and making them? Are they asking exorbitant costs, or adding crazy conditions?

If domestic R&D is necessary, countries aren't making use of their human resources efficiently. I have loads of PhD and postdoc friends with molecular or biotech experience, and they're all stuck at home crocheting and building tiki bars in garden. They'd all happily sign up to a coordinated effort to make tests, but a month after it was obviously necessary nothing has moved.

What's with the stalemate?

As far as I’m aware there are also shortages of materials needed to conduct the tests - eg chemicals needed to transport samples and reagents used in testing. It may not be so easy to set up industrial production.

Herainestold wrote: ↑

Thu Apr 16, 2020 2:32 am

Bird on a Fire wrote: ↑

Wed Apr 15, 2020 11:31 pm

I'm sure I once knew the answer to this, but I've forgotten. How come there's such a shortage of tests in Europe and the USA, when places like South Korea managed to get hold of shitloads at short notice?

https://www.cnn.com/2020/03/12/asia/cor ... index.html

This also states that the South Korean government built up testing capacity over several years. Probably due to the SARS experience.

Woodchopper wrote: ↑

Thu Apr 16, 2020 5:08 am

As far as I’m aware there are also shortages of materials needed to conduct the tests - eg chemicals needed to transport samples and reagents used in testing. It may not be so easy to set up industrial production.

Not when everybody else is trying to do it at the same time.

This is why I think it's remarkable that billions' of pounds of state of the art academic equipment, much of it taxpayer funded, is sitting idle.

It just seems an odd and widespread problem I haven't seen covered much. Instinctively I find it hard to believe that the physical act of testing is the bottleneck, but I'm very open to persuasion.

Woodchopper wrote: ↑

Thu Apr 16, 2020 5:13 am

Herainestold wrote: ↑

Thu Apr 16, 2020 2:32 am

Bird on a Fire wrote: ↑

Wed Apr 15, 2020 11:31 pm

I'm sure I once knew the answer to this, but I've forgotten. How come there's such a shortage of tests in Europe and the USA, when places like South Korea managed to get hold of shitloads at short notice?

https://www.cnn.com/2020/03/12/asia/cor ... index.html

This also states that the South Korean government built up testing capacity over several years. Probably due to the SARS experience.

Noe with link: https://www.npr.org/2020/03/12/81509781 ... rus-spread

Bird on a Fire wrote: ↑

Thu Apr 16, 2020 6:30 am

This is why I think it's remarkable that billions' of pounds of state of the art academic equipment, much of it taxpayer funded, is sitting idle.

It just seems an odd and widespread problem I haven't seen covered much. Instinctively I find it hard to believe that the physical act of testing is the bottleneck, but I'm very open to persuasion.

One problem is that the majority of these University labs will not be Containment Level 3 facilities, which is the recommended biosafety level for handling known or suspected samples. (Though the guidance is a bit more nuanced for diagnostic labs, eg some activities like RT-PCR prep can be CL2.)

https://www.gov.uk/government/publicati ... -specimens

Also, using a range of different equipment in different labs has a big risk of inconsistencies, both in terms of the actual data (eg different false negative rates?), and also stuff like turnaround times, staffing, reagent supply chains and sample tracking. What if your hospital's samples go to a Uni lab that is understaffed today because of illness, or is waiting for the qiagen kits to come in? An unexpected 24h delay in getting results back is a big deal. How do you decide which facility to send your samples to today to get the quickest results?

There is stuff going on to build capacity, but it's not an ad-hoc thrown together job. I know a couple of people who are getting seconded to this:
https://www.cam.ac.uk/news/astrazenecag ... t-covid-19

As well as validation a high-throughput method of running the PCR test, there's also a lot of work going into automating both the sample handling and the data analysis/reporting.

Report in the Guardian yesterday said that although US were aiming to develop their own test (rather than taking the Korean ones); their test ended up detecting other Coronaviruses as well, as false positives.
This meant it was unusable as a diagnostic tool. So so that is one issue that if you go down the wrong path, you can delay things.

Here in the Netherlands they've looked at blood donors, and found that about 3 percent have covid-19 antibodies. The researchers add the caveat that they are not representative of the country's population as a whole. An interesting thing they found is that younger people with antibodies have more of them than older people, which possibly ties in with the fact that younger people have a better survival rate. The research is not yet published.

Dutch news article

jaap wrote: ↑

Thu Apr 16, 2020 12:18 pm

Here in the Netherlands they've looked at blood donors, and found that about 3 percent have covid-19 antibodies. The researchers add the caveat that they are not representative of the country's population as a whole. An interesting thing they found is that younger people with antibodies have more of them than older people, which possibly ties in with the fact that younger people have a better survival rate. The research is not yet published.

Dutch news article

In what way are blood donors considered not representative of the whole population?

Presumably they are younger and healthier than average, and therefore also higher probability of asymptomatic infection, but the relevant point is whether they are representative in their chances of exposure to the virus.

AMS wrote: ↑

Thu Apr 16, 2020 9:13 am

Also, using a range of different equipment in different labs has a big risk of inconsistencies, both in terms of the actual data (eg different false negative rates?), and also stuff like turnaround times, staffing, reagent supply chains and sample tracking. What if your hospital's samples go to a Uni lab that is understaffed today because of illness, or is waiting for the qiagen kits to come in? An unexpected 24h delay in getting results back is a big deal. How do you decide which facility to send your samples to today to get the quickest results?

For sure (and thanks for the other info). But this compares with the current situation where there has been a woefully inadequate number of tests for over a month, despite the existence of one that (apparently) definitely works, and despite numerous false starts from other approaches.

A 24h delay is a much smaller deal than a 24-day delay.

Seems like letting perfect be the enemy of the good.

Aside from being younger and healthier (and therefore probably more confident in exposing themselves to the virus), blood donors, at least in England, tend to be under-representative of ethnic minorities [1], and new donors are more likely to be male and white (wrong - see post below). There's a significant mismatch between demographics of blood donors and the general population [2].

AMS wrote: ↑

Thu Apr 16, 2020 1:28 pm

jaap wrote: ↑

Thu Apr 16, 2020 12:18 pm

Here in the Netherlands they've looked at blood donors, and found that about 3 percent have covid-19 antibodies. The researchers add the caveat that they are not representative of the country's population as a whole. An interesting thing they found is that younger people with antibodies have more of them than older people, which possibly ties in with the fact that younger people have a better survival rate. The research is not yet published.

Dutch news article

In what way are blood donors considered not representative of the whole population?

Presumably they are younger and healthier than average, and therefore also higher probability of asymptomatic infection, but the relevant point is whether they are representative in their chances of exposure to the virus.

Younger healthier people are quite likely to go out more to a wider variety of indoor spaces than older sicker people, surely?

And EPD brings the facts we all need.

Actually, scrap that: most new donors are women, sorry. Some confusing language in that second reference.

Bird on a Fire wrote: ↑

Thu Apr 16, 2020 1:34 pm

AMS wrote: ↑

Thu Apr 16, 2020 9:13 am

Also, using a range of different equipment in different labs has a big risk of inconsistencies, both in terms of the actual data (eg different false negative rates?), and also stuff like turnaround times, staffing, reagent supply chains and sample tracking. What if your hospital's samples go to a Uni lab that is understaffed today because of illness, or is waiting for the qiagen kits to come in? An unexpected 24h delay in getting results back is a big deal. How do you decide which facility to send your samples to today to get the quickest results?

For sure (and thanks for the other info). But this compares with the current situation where there has been a woefully inadequate number of tests for over a month, despite the existence of one that (apparently) definitely works, and despite numerous false starts from other approaches.

A 24h delay is a much smaller deal than a 24-day delay.

Seems like letting perfect be the enemy of the good.

A major factor I've seen pointed out from someone of this or the former parish is that academic and industrial R&D labs aren't set up with equipment or staffing for processing high volumes of repetitive tests. A single hospital lab can probably process far more tests to diagnostic specs than a motley collection of university labs. The number they gave was that a single normal DGH lab usually processes 20k tests a day - not all for virology though.

Tweet

Ok, so if it's not the processing of tests, what is it? Manufacturing them? Why are people still testing out novel designs if there's one we know works?

Bird on a Fire wrote: ↑

Thu Apr 16, 2020 7:59 pm

Ok, so if it's not the processing of tests, what is it? Manufacturing them? Why are people still testing out novel designs if there's one we know works?

What people are working on (as I understand it) is how to robustly run the assays at high throughput. Lab robotics are vital to running the same assay 10,000s of times, but there are some things they don't do well. (Centrifugation, for example, or anything where visual inspection is needed, such as checking that a viscous reagent has mixed in properly.) So protocols need modification from how they are run at bench scale by a molecular biologist with a set of Gilsons.

Ah ok, so developing the automation with available kit rather than the tests themselves? (As noted above, South Korea had already sorted automation out because of SARS, and other countries failed to learn from their experiences then).