COVID19

[Marky147]

https://www.bbc.co.uk/news/uk-england-suffolk-54353697

This article struck a chord with me today. We are due to have our first child in 6 weeks time. This is very "first world problems" but we didn't get the opportunity to go to the scans together or attend any midwife appointment stuff. I am the only person allowed to be at the birth which is the most upsetting - my partner is obviously very nervous and would have felt more comforted to have her mum and family in the room or waiting close by.

No complaints of course but a very strange time to be going through this. Will be a positive end to a crap year. Generation COVID!

Little boy arrived yesterday after partner started contractions last Friday - FIVE days! They may need to stay in for a couple of nights and I can only visit between 3pm-4pm. Hoping I might be able to get them home today. Terrifying but amazing.

My missus had two huge sessions, both over 24 hours, and I can't imagine 5 days of that.   Hope Covid meant you didn't have to alternate between standing and the most uncomfortable seat in the World for 5 days.   

Hope they both get out soon and get plenty of sleep now.  As Red says they are great and completely worth it once you get through the initial few months.

I’ll take the overs.

My parents would probably say 4 decades

Approaching that mark as a parent myself, and I’m still inclined to take the overs on 4 decades..l

They most likely will, once I'm done with the 4th decade

[Doobs]

How does the actual data compare to No10’s four scenarios shown at the Saturday press conference?

The Specator is updating daily

 It doesn't look up to date on a rough eyeball.  There are 492 today and there were just shy of 400 yesterday.  Haven't they just run it to the weekend, when there are delays, and not updated when they realised what happens midweek?

My optimism a few hours ago that maybe 500 could be the worst seems to have been misplaced.  Will take a couple of weeks before the lockdown feeds through.

I'll answer this myself.  You can find the graphs here, and interestingly you can click on get the data and go through to the actual spreadsheet used to create the graphs. 

https://data.spectator.co.uk

The chart ends on Saturday 31 October, but you can just put in the data from that date into your newly created excel sheet to create a 4 Novembet version.  In that version, you can see we are pretty much going along as expected in the projections.  This is all pretty good.  Presumably we can see the effect of the partial lockdown in the future by monitoring this?

I have been playing about with this on my PC.

I am pretty sure the spectator were using actual deaths at date of death rather than reported deaths, hence when you look at Tighty's picture the downswing at the end isn't a weekend downswing but just shows the slow reporting of previous days deaths.

Looking at 7 days ago (which should allow sufficient time for most actual deaths to be included in the data), the deaths on 29/10/2020 in England were 232.  This is likely to be an increase of 30% or so on a week previously (it is 27.5% right now, but I assume a handful of cases will be added late even now).

Looking at the comparisons, that is higher than the Warwick mid projection (210) and higher than the LSHTM low projection (150), but lower than their mid projection (270).  It is lower than all 3 Imperial projections (lowest is 280).  It is also lower than the "Cambridge/PHE" projection that was out of date and has been discredited.   

Of those projections mentioned above that were close to the deaths figure of a week ago, Warwick mid has 350 cases a day by today (Wawick low has 290, LSHTM low has 250 and LSHTM has 430).  I don't think any are massively out right now (I assume we have about 300-400 deaths a day right now once the reporting lag has worked through). 

Going forward, Warwick mid has 1,000 cases a day by 28 November; LSHTM low has 1,000 cases by 5 December, Warwick mid by 19 November and LSTHM mid by 21 November.  So the 4 more reasonable projections all have 1,000 cases a day by around the end of November.  So even if we can ignore the more hysterical/out of date Cambridge PHE projection, without measures we appeared to be heading towards peak wave 1 Covid death figues around the end of this month.

Is this now going to happpen?  I don't know.  There seems to be some evidence that growth rates were slowing already (maybe because of the tier 3 areas, maybe because people were becoming more careful and maybe because of increased immunity).  Add in the lockdown from yesterday and I think it is OK to be hopeful we don't now hit 1,000 cases a day by the end of November, and we peak at a lower number. I wouldn't be in any rush to discount a third wave though.
   

Just a quick update on this as I have updated the data.

The figure for 29/10/2020 has been updated to 247, so the weekly increase back then was at least 33%.

As of 4  November (a week ago) the figure for daily deaths is now  252, but presumably there will be around 15 or 20 added to that by this time next week.

So at the time of this lockdown, deaths were still increasing by something like 15% to 20% a week.   So you could argue that the death rate was already slowing when the lockdown was announced, but it certainly hadn't peaked.  Because of the lag from infections to deaths, and the limits on testing numbers, it is hard to say if infections were still rising too.

As of a week ago, the deaths were still above the Warwick and LSHTM low projections, but were below the middle projections from all providers.  There seems a good chance right now that we peak below 1,000 deaths a day, and it may well peak below 500 a day in this wave (actuals not reported, as the reported number is already over 500).

Deaths have now totalled over 60,000 on the ONS measure.

Just a quick update on this. 

As of 11/11/2020, the deaths per day in England were 299, which was below all the forecasts.  It was 10% higher than the week previously and is still subject to possible revision. 

I have stuck to updates as of Wednesday for consisrtency.  There were 100 more deaths on Monday 9 November, though that looks a bit of an outlier.  The deaths figure on that date was significantly higher than projections.

Looking at cases data, there does seem to be some levelling off, and the cases may well be falling.  I think we need to see a few days where the UK number is below 20,000 to be sure.  Cases were 27k on 11/11/2020 vs 24k a week earlier.  All numbers after the 11th have been lower, though there were 24.5k on the 16th, so any fall isn't massively significant yet (there have only been 2 days where more than 30k cases have been reported). 

For those still sceptical of lockdowns, the Government is seemingly very lucky with the timing of their lockdowns and their easings.

I don't know what this means for Christmas gatherings, but people should be playing golf again soon (they shouldn't really have stopped in my view).       

[DaveShoelace]

Oxford trial is 70% effective after a half dose and 90% effective after two doses https://twitter.com/UniofOxford/status/1330769567409385474

[neeko]

Oxford trial is 70% effective after a half dose and 90% effective after two doses https://twitter.com/UniofOxford/status/1330769567409385474

“Today marks an important milestone in the fight against #COVID19. Interim data show the #OxfordVaccine is 70.4% effective, & tests on two dose regimens show that it could be 90%, moving us one step closer to supplying it at low cost around the world“


This week by Oxford Uni is another example of imperfect science communication, hindering the brilliant science behind it.

There were 2 parts to the study
Brazil 9000 people (2 full strength doses) 60% effective
UK 3000 people (1st half strength then full strength 2nd jab) 90% effective.

On average 70% effective.

The tweet implies that 1 jab is 60% but 2 jabs are 90%  but that is not what is happening.

Professors of Biochemistry are very good a science but should find someone else to do their tweets.

[Marky147]

This is pretty interesting.

[Marky147]

And this one just gone up, too.

[Doobs]

This is pretty interesting.

Think he is very good Marky; I like him, he doesn't seem to twist anything like a couple of other more visible COVID "experts". 

I get similar on Facebook; people from Asia/Australia just confused about why we didn't just do what they did. Freedom is surely how they are living right now and not killing masses of people for a slightly more palatable restrictions and slightly less bad hit to the economy in the short term.  Thankfully things are looking up with the vaccine and cases seem to be going down (well until we all get together at Christmas anyway).

He put one up about MMR the other day that was quite good too.  Seems if you have had MMR and spend a lot of time outside soaking in the vitamin D you have a better chance.  I am too old to have got MMR, but the lack of work opportunities has helped with the latter.

[Doobs]

How does the actual data compare to No10’s four scenarios shown at the Saturday press conference?

The Specator is updating daily

 It doesn't look up to date on a rough eyeball.  There are 492 today and there were just shy of 400 yesterday.  Haven't they just run it to the weekend, when there are delays, and not updated when they realised what happens midweek?

My optimism a few hours ago that maybe 500 could be the worst seems to have been misplaced.  Will take a couple of weeks before the lockdown feeds through.

I'll answer this myself.  You can find the graphs here, and interestingly you can click on get the data and go through to the actual spreadsheet used to create the graphs. 

https://data.spectator.co.uk

The chart ends on Saturday 31 October, but you can just put in the data from that date into your newly created excel sheet to create a 4 Novembet version.  In that version, you can see we are pretty much going along as expected in the projections.  This is all pretty good.  Presumably we can see the effect of the partial lockdown in the future by monitoring this?

I have been playing about with this on my PC.

I am pretty sure the spectator were using actual deaths at date of death rather than reported deaths, hence when you look at Tighty's picture the downswing at the end isn't a weekend downswing but just shows the slow reporting of previous days deaths.

Looking at 7 days ago (which should allow sufficient time for most actual deaths to be included in the data), the deaths on 29/10/2020 in England were 232.  This is likely to be an increase of 30% or so on a week previously (it is 27.5% right now, but I assume a handful of cases will be added late even now).

Looking at the comparisons, that is higher than the Warwick mid projection (210) and higher than the LSHTM low projection (150), but lower than their mid projection (270).  It is lower than all 3 Imperial projections (lowest is 280).  It is also lower than the "Cambridge/PHE" projection that was out of date and has been discredited.   

Of those projections mentioned above that were close to the deaths figure of a week ago, Warwick mid has 350 cases a day by today (Wawick low has 290, LSHTM low has 250 and LSHTM has 430).  I don't think any are massively out right now (I assume we have about 300-400 deaths a day right now once the reporting lag has worked through). 

Going forward, Warwick mid has 1,000 cases a day by 28 November; LSHTM low has 1,000 cases by 5 December, Warwick mid by 19 November and LSTHM mid by 21 November.  So the 4 more reasonable projections all have 1,000 cases a day by around the end of November.  So even if we can ignore the more hysterical/out of date Cambridge PHE projection, without measures we appeared to be heading towards peak wave 1 Covid death figues around the end of this month.

Is this now going to happpen?  I don't know.  There seems to be some evidence that growth rates were slowing already (maybe because of the tier 3 areas, maybe because people were becoming more careful and maybe because of increased immunity).  Add in the lockdown from yesterday and I think it is OK to be hopeful we don't now hit 1,000 cases a day by the end of November, and we peak at a lower number. I wouldn't be in any rush to discount a third wave though.
   

Just a quick update on this as I have updated the data.

The figure for 29/10/2020 has been updated to 247, so the weekly increase back then was at least 33%.

As of 4  November (a week ago) the figure for daily deaths is now  252, but presumably there will be around 15 or 20 added to that by this time next week.

So at the time of this lockdown, deaths were still increasing by something like 15% to 20% a week.   So you could argue that the death rate was already slowing when the lockdown was announced, but it certainly hadn't peaked.  Because of the lag from infections to deaths, and the limits on testing numbers, it is hard to say if infections were still rising too.

As of a week ago, the deaths were still above the Warwick and LSHTM low projections, but were below the middle projections from all providers.  There seems a good chance right now that we peak below 1,000 deaths a day, and it may well peak below 500 a day in this wave (actuals not reported, as the reported number is already over 500).

Deaths have now totalled over 60,000 on the ONS measure.

Just a quick update on this. 

As of 11/11/2020, the deaths per day in England were 299, which was below all the forecasts.  It was 10% higher than the week previously and is still subject to possible revision. 

I have stuck to updates as of Wednesday for consisrtency.  There were 100 more deaths on Monday 9 November, though that looks a bit of an outlier.  The deaths figure on that date was significantly higher than projections.

Looking at cases data, there does seem to be some levelling off, and the cases may well be falling.  I think we need to see a few days where the UK number is below 20,000 to be sure.  Cases were 27k on 11/11/2020 vs 24k a week earlier.  All numbers after the 11th have been lower, though there were 24.5k on the 16th, so any fall isn't massively significant yet (there have only been 2 days where more than 30k cases have been reported). 

For those still sceptical of lockdowns, the Government is seemingly very lucky with the timing of their lockdowns and their easings.

I don't know what this means for Christmas gatherings, but people should be playing golf again soon (they shouldn't really have stopped in my view).       

Another quick update as it is Wednesday.

Deaths were up 20% week on week on the 18 November.  That seems to be just a function of the low number the week previously, and deaths are only up about 5% a week averaging over the 7 days to that date.

On no date has the figure topped 400 a day in England and 500 a day in the UK in total (based on specimen date).  It is possible both will be slightly breached this week once the delayed reporting has worked its way through. 

The 600 death figure from yesterday is likely just a reflection of reporting delays, and I think we have likely passed the date with the highest number of deaths from this wave.  We also had a couple of days in the previous week with less than 20k cases and looking at the reported numbers from the last couple of days it seems likely that the cases are falling siginifcantly.   I'd be much happier if they waited a couple of weeks longer for the cases to decrease further before opening restaurants and pubs, but can't complain about outdoor sports and letting people spectate in smaller numbers.

FWIW I got randomly selected and my negative result took about 72 hours exactly from courier pick up.  I'd say that kind of delay seems consistent with what I see in the case reporting (most around 2 or 3 days after test). 

 

[Marky147]

Think he is very good Marky; I like him, he doesn't seem to twist anything like a couple of other more visible COVID "experts". 

I get similar on Facebook; people from Asia/Australia just confused about why we didn't just do what they did. Freedom is surely how they are living right now and not killing masses of people for a slightly more palatable restrictions and slightly less bad hit to the economy in the short term.  Thankfully things are looking up with the vaccine and cases seem to be going down (well until we all get together at Christmas anyway).

He put one up about MMR the other day that was quite good too.  Seems if you have had MMR and spend a lot of time outside soaking in the vitamin D you have a better chance.  I am too old to have got MMR, but the lack of work opportunities has helped with the latter.

Definitely good for people like me that need it explained as simply as possible. Just lays out the figures, as you say.

American's are even worse than us here, just about. I'm not doing that, don't tell me what I can do! Bla Bla Bla...

I saw that, about the correlation between having had the MMR and better outcomes if you did catch it, from what I understood?

Fingers crossed we can all get our nanobots fitted next year, and back in Vegas living the dream

[Doobs]

Think he is very good Marky; I like him, he doesn't seem to twist anything like a couple of other more visible COVID "experts". 

I get similar on Facebook; people from Asia/Australia just confused about why we didn't just do what they did. Freedom is surely how they are living right now and not killing masses of people for a slightly more palatable restrictions and slightly less bad hit to the economy in the short term.  Thankfully things are looking up with the vaccine and cases seem to be going down (well until we all get together at Christmas anyway).

He put one up about MMR the other day that was quite good too.  Seems if you have had MMR and spend a lot of time outside soaking in the vitamin D you have a better chance.  I am too old to have got MMR, but the lack of work opportunities has helped with the latter.

Definitely good for people like me that need it explained as simply as possible. Just lays out the figures, as you say.

American's are even worse than us here, just about. I'm not doing that, don't tell me what I can do! Bla Bla Bla...

I saw that, about the correlation between having had the MMR and better outcomes if you did catch it, from what I understood?

Fingers crossed we can all get our nanobots fitted next year, and back in Vegas living the dream

Yep, that was it.  That fraudster, ex-Dr Wakefield managing to do lasting damage in two ways. 

Much of the virus scepticism stems from him spreading lies for money on MMR; and all those that missed MMR now have increased risk of a worse Covid experience.   

Be careful with the nanobots Marky, THEY are out to control us.

[Marky147]

Plenty of charlatans nicking a good lump from all this, as most things down the years, no doubt. Easier to do now, as the world is a much smaller place, and so easy to capture vast audiences.

Can inject me with whatever trackers they want. Won't glean much information of interest from me. I go to the gym, the hospital, and the KFC drive-thru.

[Doobs]

Talking of charlatans, I enjoyed this exchange on the "statistics" guy's fake chart that the Daily Mail reproduced (no surprise).

https://twitter.com/john_actuary/status/1329145198677528576

And this from a right wing Tory MP

https://twitter.com/NeilDotObrien/status/1330830800850472961

There was another exchange he had not long ago where he showed that he really didn't get exponential; which is a fairly common thing to misunderstand, but it is a really weird thing for a mathematician to not understand. 

Note, there are plenty of real statistics guys to follow on twitter.

This should be interesting, but might well be just for stats nerds (not Jon obv.)

https://www.linkedin.com/events/modellingcovid-19-whathavewelea6734137085323808768/

[Doobs]

Talking of charlatans, I enjoyed this exchange on the "statistics" guy's fake chart that the Daily Mail reproduced (no surprise).

https://twitter.com/john_actuary/status/1329145198677528576

And this from a right wing Tory MP

https://twitter.com/NeilDotObrien/status/1330830800850472961

There was another exchange he had not long ago where he showed that he really didn't get exponential; which is a fairly common thing to misunderstand, but it is a really weird thing for a mathematician to not understand. 

Note, there are plenty of real statistics guys to follow on twitter.

This should be interesting, but might well be just for stats nerds (not Jon obv.)

https://www.linkedin.com/events/modellingcovid-19-whathavewelea6734137085323808768/

The actuarial webinar from yesterday is on youtube. 



There is a lot of info in there, as it picks from 6 months of research.

I might go through it again at some stage and link to snippets (eg the bit where mask wearing efficiency is discussed).

I think the ordering is a bit wrong for the general public, as the more accessible/interesting stuff is towards the end.

Things I found particularly interesting:

There is a bit in the international part where South Africa is discussed by a South African actuary and given the excess deaths treatment.  In that part it is proposed that a significant chunk of the population has been infected and as much as half the population could be infected by the end of the year (it is about 55 minutes in).

I think Stuart McDonald is very good and his section seems accessible on communicating model results starts at 59 minutes.  He goes through some of the more notorious bad communications of the epidemic (think he is too kind to Sunetra Gupta who seems to have more than doubled down and gone along with some bad interpretations of her study) .   

Stuart McDonald has a twitter https://twitter.com/ActuaryByDay, which seems very good.  I have briefly come across him in the past and am not surprised that he comes across well.

Comorbidities get a mention at about 1 hour 12 minutes in and covers the "they were about to die anyway" myth.  You can also see how wrong this is by looking at the earlier section (about 20 minutes in), where the excess deaths are shown over the year (there really hasn't been significantly better mortality across once the "dry tinder" from March had gone).

One of David Spiegelhalter's charts gets a mention too, and he is normally very good https://twitter.com/d_spiegel.

I will leave it for now, as have to get something else done

[Marky147]

One of my mates does nothing but crow about masks/lockdowns, and one of his mates posted some link to a video by Gupta, and I watched some of it, but it was from 6m ago. The comments were basically 'this has aged well', and it was something to do with lockdown sceptics, too.
 
Will send him that to look at, to send to his friend

[Marky147]

Watched the last half hour, and interesting stuff. Should be mandatory for all to watch.

Will try and watch the first hour tonight, and then wait for your breakdown for donks so I understand it