Daily numbers, graphs, analysis thread 12

[BigChocFrenzy]

Welcome to thread 12 of the daily updates

Resource links:

Slides & data UK govt pressers
UK dashboard sub-national data, local authorities
Beta Uk dashboard deaths, cases, hospitals, tests, partially sub-national
UK stats updated daily by PHE & DHSC
ONS UK statistics for CV related deaths, released weekly each Tuesday
PHE surveillance report infections & deaths released every Thursday with sep. infographic
NHS England stats including breakdown by Hospital Trust
FT Daily updates
HSJ Healthcare updates
Worldometer UK page
Plot FT graphs compare countries deaths, cases / million pop. / log / linear
Covidly.com filter graphs compare countries
Plot COVID Graphs Our World in Data

We welcome factual, data driven, and civil discussions from all contributors 📈📶👍

Thank you for the new thread🙏 and it's good to see pubs behaving responsibly.
Surely situations like this could be a big part of track and trace, ie if all establishments have to log contact details of everyone who enters them then this could provide a valuable mechanism for tracking and containing cases of the virus?

Thanks BigChoc

Re: ONS figures. If we are now below the 5 year average, but the total deaths still include a high proportion of Covid deaths, does it follow that if we didn't have Covid, we would be WAY below the 5 year average? This would seem to add to the view that those that have recently died of, or with, Covid, may have died this year anyway, but of something else. It is going to be quite interesting to see the end of year results, and whether our excess deaths will actually disappear on a longer rolling basis. The rub being, sadly, that we will then have many other early deaths of those with cancer or other diseases, who may have survived longer without the Covid crises. Ugh, such sad times.

@Baaaahhhhh

Re: ONS figures. If we are now below the 5 year average, but the total deaths still include a high proportion of Covid deaths, does it follow that if we didn't have Covid, we would be WAY below the 5 year average? This would seem to add to the view that those that have recently died of, or with, Covid, may have died this year anyway, but of something else. It is going to be quite interesting to see the end of year results, and whether our excess deaths will actually disappear on a longer rolling basis. The rub being, sadly, that we will then have many other early deaths of those with cancer or other diseases, who may have survived longer without the Covid crises. Ugh, such sad times.

Basically, yes. Which is what the ONS have actually said today - that we may be entering a period in which death numbers actually fall below five year averages. We've just had 2 consecutives week in June so far where that has been the case. From their report -

" Some of these deaths would have likely occurred over the duration of the year but have occurred earlier because of the coronavirus. These deaths occurring earlier than expected could mean we see start to see a period of deaths below the five-year average."

That's interesting because I think it's the first time they have actually suggested that, though it's been suggested by a number of others before, given the age and health profile of a significant amount of deaths. It's a contentious issue as it can be construed as saying "it doesn't matter as they would have died soon anyway."

It's also why looking at "excess deaths" as a comparator is only interesting in the long term. We need to see where we are in a year's time, not now.

Coronavirus: Majority testing positive have no symptoms www.bbc.co.uk/news/health-53320155

So am I right and this is way higher on the asymtamatic cases than other estimates (80 percent) previously I thought we were working on about 40 percent?

It would seem to have quite serious issues re moving forward if so, we would need regular testing well beyond just health care workers and care home residents if we are to really keep the virus suppressed.

@Qasd

The figure is for those who aren’t showing symptoms on the day of their test. The last sentence in the article states: “Some people testing positive without symptoms might go on to develop symptoms, or they may have already had symptoms and cleared them.”....

The report says-

"While those who have symptoms are more likely to test positive than those without symptoms, out of those within our study who have ever tested positive for COVID-19, 33% reported any evidence of symptoms around the time of their positive swab test."

• so 67% did have symptoms, which suggests some developed them just after the test (or subsequent tests as some had more than one).It also includes people who didn't answer the question...

Also -"There is evidence to suggest that infection rates are higher among people who have reported coming into recent contact with a known case of the coronavirus than those who have had no reported contact with potential cases" - which suggests (possibly) the lower chance of being infected by someone who is asymptomatic, possibly?

Very low numbers involved though -the study only has 115 individuals who tested positive so any conclusions should be VERY tentative.

Baaahhh and Patricia, the idea that “these deaths would have occurred anyway” is only true for those who, pre-covid, had very, very low life expectancy. These will be a very small proportion of the total deaths (eg. Those with terminal cancer rather than those who had diabetes, for example)

While measuring by excess deaths is better than other methods, it is still actually quite crude because it ignores a number of key things - including changes in the size and age profile of the population, but also underlying trends in life expectancy over the relevant period. The UK has had a long, sustained period of improvements in life expectancy, which slowed somewhat over the period 2013 to 2018, but had appeared to pick up again in 2019 (and that trend continued pre-Covid in 2020). Taking this into account, it would be reasonable to have expected lower deaths now than the 5 year average as a starting point before factoring in Covid.

Qasd, the number of asymptomatic cases does appear high. I wonder if it is distorted by the fact that the survey excludes those in hospital; someone with severe symptoms isn’t included because they are no longer at home but in hospital instead. Put another way, it shows proportions of people asympyomatic and with mild disease, but excludes those severe enough to need medical attention.

And final post on today’s ONS data, which highlights the time lag between the daily reported deaths and actual deaths that week. ONS report Covid deaths for the relevant week were 650, compared with around 800 reported deaths from the daily reporting for same week.

hi everyone! I was reading the story about someone testing positive after going to a pub in Somerset but in all 4 areas on Somerset the data shows no new cases since the end of June? I just wondered if all positive tests are recorded in the same place?

@AprilLady - the idea that “these deaths would have occurred anyway” is only true for those who, pre-covid, had very, very low life expectancy. These will be a very small proportion of the total deaths (eg. Those with terminal cancer rather than those who had diabetes, for example)
I'm not sure about that, to be honest. I think it's quite possible, for example, that there were a fair number of people who might well have succumbed to flu this coming winter who have passed from Covid instead. But we aren't going to know that of course until after flu season.

Oh and yes the time lag is interesting - this is from @cricketwyvern on twitter, plotting England ONS death by day of death vs England hospital deaths by day of death. Showing that England hospital deaths have actually been below 50 a day for some time, and that this week all settings deaths probably will be. Although it won't look like that because of the lags reporting - we might be in for largish numbers today and tomorrow because of the weekend lag.

@Cusano34

hi everyone! I was reading the story about someone testing positive after going to a pub in Somerset but in all 4 areas on Somerset the data shows no new cases since the end of June? I just wondered if all positive tests are recorded in the same place?

Hi there - the most likely scenario is that a test taken on, say Sunday, may well not have been registered yet. It would only have been flagged as a positive yesterday, so won't have hit the stats yet.

It's also possible that the person doesn't live in Somerset!

@PatriciaHolm yes all those points make absolute sense, thanks! :)

Placemarking

Thanks for the new thread.

Is there any early analysis of Victoria?

The reduction below the average in deaths could be due to other things as well as some covid-19 deaths bringing deaths forward e.g. less sporting activity, fewer people on the roads (I'd guess this is a big factor), fewer accidents on nights out etc etc.

There is this study which suggests mean years of life lost from covid deaths are 11 for women and 13 for men taking into account underlying conditions. wellcomeopenresearch.org/articles/5-75

I think the narrative - which is very convenient for the government - that all covid deaths are deaths brought forward is pretty much flat out wrong. There was a 'more or less' episode (radio 4 -excellent on statistics etc) which addressed this too and said the same thing.

@NeurotrashWarrior

Thanks for the new thread.

Is there any early analysis of Victoria?

Attached is the infection graph. Victoria would appear to have 35 people in hospital, up from 22 on Friday www.health.gov.au/news/health-alerts/novel-coronavirus-2019-ncov-health-alert/coronavirus-covid-19-current-situation-and-case-numbers#cases-admitted-to-hospitals

Victoria issue a daily report and press release -
app.powerbi.com/view?r=eyJrIjoiODBmMmE3NWQtZWNlNC00OWRkLTk1NjYtMjM2YTY1MjI2NzdjIiwidCI6ImMwZTA2MDFmLTBmYWMtNDQ5Yy05Yzg4LWExMDRjNGViOWYyOCJ9

Cases seem focused on 3 suburbs as well as Melbourne. Average age of confirmed case (over all cases over the last 5 months) 37.

They seem to have some idea where cases are - lots of family clusters:

• Of the new cases which have already been linked to outbreaks, the breakdown includes: -13 cases relating to the North Melbourne and Flemington public housing towers, with the total now 69. -12 new cases linked to the Al-Taqwa College outbreak, with the total now 90. -4 new cases have been linked to the Northern Hospital in Epping, with the total now 9. This is made up of 8 staff and 1 household contact.

• 1 case linked to Aitken Hill Primary School in Craigieburn, with the total now 10. The case is a household contact of a confirmed case.
• The remaining new cases are linked to existing family clusters in Truganina, Patterson Lakes/Lysterfield, Fawkner and Sunshine West.*

@Langismyhero

The reduction below the average in deaths could be due to other things as well as some covid-19 deaths bringing deaths forward e.g. less sporting activity, fewer people on the roads (I'd guess this is a big factor), fewer accidents on nights out etc etc.

There is this study which suggests mean years of life lost from covid deaths are 11 for women and 13 for men taking into account underlying conditions. wellcomeopenresearch.org/articles/5-75

I think the narrative - which is very convenient for the government - that all covid deaths are deaths brought forward is pretty much flat out wrong. There was a 'more or less' episode (radio 4 -excellent on statistics etc) which addressed this too and said the same thing.

Indeed. I don't think anyone (here, at least) is claiming that anything like all deaths are deaths brought forward. Some will be though.

I hope someone will do an update to that study - it dates from April, and seems to be based on a dataset of 701 Italians who died in March. There must be considerably more data out there now.

*The reduction below the average in deaths could be due to other things as well as some covid-19 deaths bringing deaths forward e.g. less sporting activity, fewer people on the roads (I'd guess this is a big factor), fewer accidents on nights out etc etc. "

1,521 people died on the UK roads in 2018. Reducing that by, say, a third, isn't going to have much of difference. (I was surprised by how few it was tbh).

If 2018 (i think) was a particularly bad year for flu and this last winter wasn't, wouldn't we expect deaths to be below the 5 year rolling average, although I suppose by June that wouldn''t be having such an effect?

Thanks Patricia. Someone elsewhere who lives there said 1000s refused to get tested. Very selfish as now the schools are back to online teaching.

A question that has likely been covered already - why does UK, Spain, Italy, etc have such a high death:case ratio? The Americas, and others, seem to have much more positive outcomes. Is there some underreporting?

A very meaningful proportion of care home deaths will be deaths brought forward. We discussed it here several threads ago.

The average length of care home stay is estimated at 26 to 30 months depending on the source. It means that on average the residual life expectancy for an average care home resident is half of this, or 13 to 15 months.

The topic of care home deaths is highly emotional and gets a lot of attention but from the statistical perspective these are indeed deaths brought forward.

The conclusion for deaths in community (especially for middle aged population) will be very different.

Stats for today -

581 cases
155 deaths (most non hospital - UK hospital today was 36).

Last Monday - 689 cases, 155 deaths.