Coronavirus

[johnb]

Some of us are using the Zoe/King's College, London Covid-19 app.

There has been an update of the results which suggests that, after analysing the data, the number of symptomatic Covid-19 cases has fallen by 71% in the last two weeks.

https://covid.joinzoe.com/data

[Bryn]

Devastated to learn from John Tilbury that our dear friend of over 50 years, Carole Finer/Chant died this morning from Covid-19. I first got to know her as a fellow member of Cornelius Cardew's Experimental Music Class in 1969. In recent years she presented Sound Out on Resonance FM. She will be very sorely missed by her Scratch Orchestra colleagues and the banjo-playing community. It was only the weekend before last that I stayed at her place for Steve Beresford's 70th birthday celebrations at Cafe OTO. You will be remembered with love, Carole.

(My webcam was playing up, so I remain 'faceless'.)

[Dave2002]

I am wondering how well the progress of the pandemic is being modelled, and closely the observed measured values match up to any models.

The SIR model, which is one plausible baseline model doesn't seem to match what we know or are told very well - unless there are lots of hidden and still undetected cases. In that model, if one infected individual is "dropped" into a population, then providing he or she mixes well and has a lot of contact with others (what is called by some the reproduction number - and mentioned in the press conferencees) so that the reproduction number is greater than 1, perhaps 1.5 or more, then there is an onset of an exponentially rising number of cases. If the period for a doubling of cases is just a few days (say 3-6 - which seems to be what was observed a few weeks ago), then if that carried on for about 26 periods (say between 3*26 b= 78 to 6*26=156 days) then virtually everyone in a country the size of the UK would have it in 4-5 months.

CV-19 is a nasty disease and kills a significant proportion of people infected by it, while others only experience mild symptoms, and some seem not to notice any symptoms at all.

We don't know when the first case came into the UK, and I am not sure if anyone does know at the moment. There was mention in a newspaper of cases being detected at the end of January - https://www.manchestereveningnews.co...es-uk-17663843

The progress of the disease has been rapid and people have died, but - and this is perhaps really scary - very many more people could have died, and perhaps should have done according to models such as SIR.

What should happen, according to models like SIR, is that the number of susceptible individuals in the population drops as the number of infections increases. Also, more people "recover" (which in the terms of the model, means that they either do recover, and presumably become at least for a while immune, and non-infectious, or they die and are taken out of the population). Eventually there is a sort of equilibrium, after which there are very few new cases of actively infected/infectious people. However, even if (heaven forbid) one million people become infected and the case fatality rate stays at about 10%, leading to 100,000 deaths, the effects on the proportion of susceptible people in the whole UK population, and also on the proportion of immune/previously infected people, would seem to be relatively insignificant. Thus even after much carnage, and a great deal of work by medical workers, the eventual situation is still pretty much the same as before - virtually the whole population is still at risk.

If there had been an extremely large number of unobserved cases and asymptomatic people, then the proportions in the Susceptible and Recovered groups would have changed significantly, and the probability of further outbreaks and a large number of future infections and deaths would be reduced, though if the proportion of the Susceptible group had dropped to around 50% of the total population, then there would still be the possiblity of future waves of the same or a very similar disease in future years. However, if the virus did not mutate unfavourably, over a number of years something like herd immunity could build up, and the virus would no longer be a major threat.

This does not appear to be what has happened. The number of Susceptible people in our population is, for almost all practical purposes, based on the data we have been told about, almost exactly the same as it was at the start of the outbreak in the UK. In Iceland (which may be significantly different from the UK) measurements of the asymptomatic cases have been made, and estimated at about 6% of the population. If similar proportions are applied to the UK, this again makes very little difference to the overall picture of susceptibility.

I would love to be proved wrong about this, or for someone to tell me that my arguments are incorrect - and explain why, but based on what I know at present, that is my current take on it all.

[Serial_Apologist]

(My webcam was playing up, so I remain 'faceless'.)

It made me think of Stockhausen's idea of a music played by an orchestra of performers, each one located on separate planets.

[LezLee]

This might be of interest:


‘Over 100 patients taking part in clinical trial for COVID-19 treatment at NHS Greater Glasgow and Clyde

Patients in Glasgow have started receiving potential treatment for COVID-19 as part of a clinical trial at NHS Greater Glasgow and Clyde, with 120 patients recruited so far. The RECOVERY clinical trial is one of the fastest growing trials looking at potential treatment of COVID-19.

NHS Greater Glasgow and Clyde anaesthetists, respiratory physicians, emergency medicine consultants, pharmacy teams and nurses are leading the board’s involvement in this UK-wide study, funded by the UK government.

The RECOVERY trial aims to find effective treatment for coronavirus patients by testing pre-existing medicine. This includes steroids, antivirals and antimalarial agents. Antibiotics may also be added at a future amendment.

The trial is adaptive in design so treatments that show promise will be used more frequently.

All patients with COVID-19 that are treated at Queen Elizabeth University Hospital, Glasgow Royal Infirmary, IRH and RAH will be offered to take part in the clinical trial. It is a randomised-controlled trial where patients are given an active drug or standard care.

This is part of NHS Greater Glasgow and Clyde’s efforts to move forward research and development related to COVID-19 as fast as possible.

Dr Jennifer Armstrong, Medical Director of NHS Greater Glasgow and Clyde, welcomed the move and said:

“Our teams have been incredible in rising up to the challenge of COVID-19. Not just in the care and treatment of patients but also their dedication to improving our knowledge of the virus through clinical trials. This means our patients are receiving the most up to date treatment available.”

Professor Julie Brittenden, Research and Development Director, NHS Greater Glasgow and Clyde, said:

“I want to thank all of our staff for their efforts during this incredibly challenging time. Our research, development and innovation teams are working hard to improve the diagnosis and treatment of COVID-19.

“We’ve been able to rapidly mobilise our teams to take part in clinical trials like RECOVERY and this is a testament to their dedication to the cause.”

Dr Kathryn Puxty, Intensive Care Consultant at Glasgow Royal Infirmary, said:

“As we care for more and more patients with COVID-19, we are seeing first-hand the need for effective treatment.

“We will be asking patients with COVID-19 who we treat in hospital, if they want to be involved in this study in the hope that we can improve care as quickly as possible.”

Working with partners across UK, Scotland is leading, enabling and delivering world-class COVID-19 research as part of coordinated efforts to gather reliable evidence. Facilitated through the Chief Scientist Office (CSO) of Scottish Government and NHS Research Scotland, a single, national prioritisation process for COVID-19 research draws on expert advice across the UK. This prioritises studies which hold the most potential, prevents duplication of effort and ensures the resources and capacity of the health care system are not exceeded.

Charles Weller, General Manager, NHS Research Scotland comments: “RECOVERY has been the fastest growing clinical trial in medical history; and a crucial part of our efforts to better understand and tackle COVID-19. I want to thank all teams for their commitment and professionalism to this national priority study.

ENDS

[johnb]

I am wondering how well the progress of the pandemic is being modelled, and closely the observed measured values match up to any models.
.....
However, even if (heaven forbid) one million people become infected and the case fatality rate stays at about 10%, leading to 100,000 deaths, the effects on the proportion of susceptible people in the whole UK population, and also on the proportion of immune/previously infected people, would seem to be relatively insignificant.

I haven't digested your post as yet but saw you are using the "Case Fatality Ratio" (CFR) instead of the "Infection Fatality Ratio" (IFR).

My understanding is that the CFR is the ratio of the number of deaths to the number of people diagnosed with the disease (i.e. confirmed cases). This figure is greatly dependent on the level of testing, e.g. I seem to remember that the CFR for the UK, at somewhere around 13%, is very high in comparison to many other countries. This is because the UK has had a low level of testing and, until relatively recently, has concentrated on only testing those people who are ill enough to warrant admission to hospital.

The Infection Fatality Ratio is the ratio of the number of deaths to the number of people infected (once again - my understanding). This is probably what you need to use for your model, rather than the CFR.

Estimates of the IFR vary. A figure of "between 0.5% and 1%" is often used and a recent study concluded that the IFR in China was probably around 0.66% (within low/high of 0.39% to 1.33%).

However, even if (heaven forbid) one million people become infected and the case fatality rate stays at about 10%, leading to 100,000 deaths.....

As at yesterday we had 13,729 hospital deaths. That alone would suggest over one million people had been infected, say, three or four weeks ago.

Not sure this has any bearing on your question though.

[Pulcinella]

This might be of interest:
......
ENDS

Not just in Scotland:

https://www.theguardian.com/world/2020/apr/17/world-biggest-drug-trial-covid-19-uk

[Pulcinella]

Should we all be folding@home to help the research effort?

https://foldingathome.org/

https://www.theguardian.com/technolo...at-coronavirus

Many thanks for the alert. Can/should we establish a team?

My first task completed!
Bryn: Did you choose a username and create a team, or did you just stay Anonymous and part of team 0, as I did?
I'd be happy to join a team (FoR3?) but didn't see how you created one, and I may well have lost any 'personal' points from that first task now, as they will probably already have been assigned to Team 0.
No worries; it's the fact that my spare CPU capacity might be doing something useful that matters.

[Bryn]

I joined the Atheists et al group using an old BBC Forum name which Alison still addresses me by. 16 work units under my belt (17 by 6 pm). That's using two i7 based laptops at full blast on a 106 Meg Internet connection.

[Pulcinella]

I joined the Atheists et al group using an old BBC Forum name which Alison still addresses me by. 16 work units under my belt (17 by 6 pm). That's using two i7 based laptops at full blast on a 106 Meg Internet connection.

Does it keep a record of individual contributions then, if you give yourself a name?
Current work unit (maybe it has done more than I know about) has been assigned 4.87 days on my computer, but has an ETA of 2 h 38 min as I type (already 23% done)!

PS: Ah! Spotted the Stats, Teams, and Usernames page in FAQ.

[Bryn]

Does it keep a record of individual contributions then, if you give yourself a name?
Current work unit (maybe it has done more than I know about) has been assigned 4.87 days on my computer, but has an ETA of 2 h 38 min as I type (already 23% done)!

PS: Ah! Spotted the Stats, Teams, and Usernames page in FAQ.

I didn't want to spoil your joy of discovery.

I have not bothered to look into what the points system is all about but they are accumulating very quickly for the group. I don't know what to do with those not transferred when I opted to join a group.

[Dave2002]

As at yesterday we had 13,729 hospital deaths. That alone would suggest over one million people had been infected, say, three or four weeks ago.

Not sure this has any bearing on your question though.

Thanks for the clarification re terminology - CFR vs IFR.

One hint you've given is that because some of the data is heavily lagged, that the actual infections may be more in line with the SIR model than I thought. However, this is unlikely to follow the patterns of the worst pandemic trajectories, in which a very significant proportion of a population (susceptibles) becomes infected and recovers, thus significantly reducing the number of susceptible people in the population. The implication of this, as I see it, is that the proportion of susceptible people in the population hardly changes, so that advisers like Sir Patrick Valance are right to suggest that it would be a really bad strategy to let up on the clamp down on this virus now, as it could easily flare up again, in almost exactly the same way.

Opening up borders, relaxing constraints on long distance travel, and other policies without appropriate precautions, could allow carriers of the virus back in, to reinfect others.

There may well be second and third waves, but hopefully (??) the country will have developed newer strategies, and be in a better position to cope. Testing and case tracking would seem to be essential until such time as either the virus weakens, or a vaccine and other therapeutic treatments are found and put into operation. Releasing the lock down in a gung ho way will very likely put the country right back to square one.

[johnb]

You can get a back of the envelope estimate of the number of people likely to be infected in the first wave.

Today, one of the people addressing Jeremy Hunt's Select Committee said the deaths during the first wave could amount to 40k. This was a "off the top of my head" type of figure but isn't totally unreasonable. The current number of hospital deaths is now approaching 15k and we think the deaths are starting to plateau (the hospital bed figures* seem to indicate that we *might* have just past the peak). The descent from the plateau is likely to be slow. Then there are the non-hospital deaths to add in.

So if, for the sake of argument, we take 40k and use a IFR of between 1% and 0.5% that gives between 4m and 8m infections - roughly between 6% and 12% of the population. No where near the often quoted figure of 60% for herd immunity.

* hospital bed figures
If anyone is interested in these they are available in the datasets (Excel files) that accompany the daily press briefings. The latest one being available here: https://www.gov.uk/government/public...-17-april-2020

You need to create a pivot table to knock the data into some form of sensible shape:

- on the Hospital beds sheet - highlight the data (including the column headings)
- Click on Pivot Table option (in my OLD Excel it is under the Data menu)
- Next - Next - (New Sheet) Finish
- A blank pivot table will be created and a box will appear showing a list of fields to drag to the row headings, column headings and data area.
- Drag "Day of Date" to the column on the left (the row headings)
- Drag "Hospital Beds" to the main area ("Drag Data Items Here")
- Drag "Region" to the top of the total column that has appeared (they will be column headings)

... and the raw data is miraculously converted into a useful table.

[Dave2002]

... and the raw data is miraculously converted into a useful table.

There does seem - even behind the scenes - to be some convergence on infection estimates which are not going to support a "herd immunity" approach to this.

I expect that if more data processing is needed to highlight new points, that someone will do it. I haven't used pivot tables for a while - it was interesting - up to a point - to see if I could still figure out how to do them.

[MrGongGong]

We don't need to worry about collaborating with the EU anymore when we have these fine folks as our friends

Let's see how long till the Kippers/exkippers start on the same nonsense here shall we?
Though as they are all now in Spalding doing the harvest they are probably too tired to do anything.........

https://www.theguardian.com/us-news/2020/apr/17/trump-liberate-tweets-coronavirus-stay-at-home-orders