Daily numbers, graphs, analysis thread 7

[Barracker]

Welcome to thread 7 of the daily updates.

Resource links:
Worldometer UK page
Financial Times Daily updates and graphs
HSJ Coronavirus updates
Johns Hopkins Coronavirus Resource Centre
NHS England stats, including breakdown by Hospital Trust
Covidly.com to filter graphs using selected data filters
ONS statistics for CV related deaths outside hospitals, released weekly each Tuesday

Thank you to all contributors for their factual, data driven, and civil discussions.

I just wonder what causes a low R0 in a densely populated city, even allowing for higher initial key worker infection.

As I posted, I believe R0 is calculated from the number of new cases over a period of days and the change in that number over those days

London cases falling sharply would give a low R0

Glass that chart you put up seems imply ex smokers are at greater risk than current smokers?

I've also spent some time today looking at the fact infection rate positives have remained fairly stable although we are reliably testing 65-70k people now compared to 10-15m a month ago.
I've been trying to work out the percentage of those tested who are positive (currently 15% compared to 25-50% when tests were lower.

What I then tried to work out and got stuck on (5 mile walk today in sun fried my brain!) is what we could estimate the number of positive cases we'd have had using this data if your tests had been higher at that time.

So, given we have minimal idea regarding the number of cases (and increase affected by pattern of testing) how does the Govt estimate R? Anyone know?

So, given we have minimal idea regarding the number of cases (and increase affected by pattern of testing) how does the Govt estimate R? Anyone know?

Indeed, we need to have a clear idea of the number of cases at a set date in the past, and the number of new cases. Do we have this? Hospitalisation numbers might be a decent proxy?... though that wouldn’t work for care homes.

Also, it would be very useful to know the R0 in care homes and outside.

For instance, for simplicity say new infections were split 50:50 between care homes and outside care homes, if the R0 in care homes was, say, 3, and the R0 outside care homes was very low at 0.4, the total R0 would be 1.7.

Taking the total of 1.7, it would appear that we would need to double down on the lock down, when in actual fact wider society could open up a lot more.

The same applies with my hypothesis on London’s low R0, but in reverse... If it’s lower R0 is possible due to relatively high levels of immunity amongst those population segments who have been more exposed to infection, as soon as they are exposed to those population segments who haven’t been exposed, the ability for Covid to spread increases, causing R0 to rise faster than it would have if infections had been evenly distributed.

The ONS study has adjusted for 'region'. I find this a little hard to countenance as there is no 'regional' covid-19 except possibly in London. It should be adjusted on an individual local basis, as Luton has no relation with North Norfolk.

I read this just now on the BBC website:

“Health services would be overwhelmed with the sheer number of people requiring hospitalisation, the WHO warned. An estimated 3.6-5.5 million people could need hospital treatment for the virus. Of those, around 52,000-107,000 would need breathing support.”

So just 2% or so of those requiring hospital treatment would need breathing support? Makes no sense whatsoever.

Sorry to have been AWOL. Some real life stuff demanding my attention.
I'll try to catch up with the daily updates tomorrow.

https://news.sky.com/story/key-r-rate-of-measuring-covid-19-spread-rises-due-to-care-home-infections-say-experts-11984906

Sir Ian Diamond, the Office for National Statistics chief, said he agreed with earlier comments by scientist Professor John Edmunds who claimed it has risen from around 0.6 a few weeks ago because of the rapid spread in nursing homes.

"That is driven by the epidemic in care homes, he would say and I would not demur from that,"
Sir Ian told the Downing Street COVID-19 news conference.

Professor Edmunds earlier told MPs on the Science and Technology Committee that
the R value is currently between 0.75 and one.

Peak in care home deaths was 7-9 days after hospital death peak

Faisal Islamm@faisalislam*

Some background factors to bear in mind for comparing UK and Italy...

Population? UK 66m vs Italy 60m

BUT Demographics?
Italy older than UK, so over 65 population
UK: 12.2m Italy: 13.8m

Obesity?
UK 20% of population. Italy 11%

< Being old seems a much higher risk than obesity >

Glass interesting that asthma isn't such a risk factor for dying. The only thing is though that asthma can be exacerbated and therefore be a nastier illness than in non asthmatics. From that chart the factor for death in this virus is more to do with inflammation in the body/ organs which the other risk factors listed eg diabetes etc seem to have more risk factor for.

(My personal fear for catching it with asthma is battling through with an over active reaction in lungs and how long it takes to calm down, and the associated exhaustion, lasting weeks. As that's what happens in most viruses with my asthma.)

@BigChocFrenzy some people on here seem to be cackling with glee that the fatties are going to be meeting a self-induced early death. However when one adjusts for age (most people who are not fat are young and essentially have no risk of dying), then it disappears or exists as a tiny tiny factor, except possibly for the morbidly obese group, where here there is a reason for the 'morbidly' adjective because yes it is killing you

There is a straightforward exponentially increasing risk with age, and a substantially greater risk for men than women of all ages. These are biological facts. 60% of UK deaths are 80+. Countries with no substantial 80+ population - consider India with 378 million people in 1939 Vs the UK with 48 million and compare to present day.

There is a greater risk for say black people in the UK, but we don't know if this is a result of say employment differences, cultural differences, etc., or has some biological basis.

The massive Ben Goldacre study whose chart is linked above seems to have controlled for sociological and also inclination-to-certain-illnesses within the bame community and still concludes just being bame is a significant risk factor. I don't understand enough about these studies to be sure I am drawing the right conclusions what do you smart people conclude?

mobile.twitter.com/bengoldacre/status/1258372975004389379

weird, gerbil if one assumes that the % of infected people who die, or meet hospitalization criteria remains stable (e.g. 0.5% infected people die, as an entirely made up stat - it doesnt really matter) then one can calculate R0 by looking at how much that is increasing/decreasing
A fall in hospitalizations means R0 1. The scale of the drop/increase tells you how far above/below you are.

SHoots I have been seeing this disgraceful victim-blaming on several MN threads,
which I agree is not backed up by the stats

When we look at the hospital ICU tables, the BMI percentages seem to roughly match the general population at those ages

Also, I've posted elsewhere, Germany is very close behind the UK in the % obesity, so it didn't damage their "league rating"

There are so many threads where women of high BMI are terrified they will die, that they won't even receive treatment
They have long been primed to feel guilty about weight
However, statistically their risk is small, especially as most seem young anyway - so at v low risk

What has surprised me are the stats on smokers, another of the many aspects that need to be researched

Over future decades, this pandemic will be producing many thousands of papers and PhD theses over several fields

I was looking at some Scottish stats linked on here (not sure if this thread, sorry) where the deaths split by sex was pretty much 50:50. I thought this was interesting but haven't heard any comment on why Scotland might be different to the general picture. Has anyone read anything on this?

What is the 'hazard' in the chart in @GlassOfProsecco's post?

Is it infection, admission, ICU admission or death?

Wouldn't a hazard ratio of 2 indicate that there were twice as many cases in the risk factor group?

University of Antwerp study of around 3,400 blood samples show around 6% of Belgians with antibodies now. The first study two weeks ago showed that 3% of Belgians had antibodies. A different study by the Red Cross published last week showed that 4.5% of Belgians had antibodies.

Peak of infections here was April 10th, deaths peaked on April 12th

Sorry I'm at work just now so can't answer fully, but here's more detailed info;

opensafely.org/outputs/2020/05/covid-risk-factors/

I report most days on the app and today have been asked whether I still have periods as they are looking at links between COVID and menopause.