Daily numbers, graphs, analysis thread 6

[Barracker]

Welcome to thread 6 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.

It concerns me that many of those minimising the importance of elderly deaths
are the same people who are eagerly pushing to end lockdown

That's the Toby Young crew, the Social Darwinists

We need to be very careful about just "othering" and writing off the elderly and vulnerable as expendable,
just so we can get back to normal, restart the economy

@Puzzledandpissedoff

I accept that being in a cruise ship brings people into close proximity, but not any more than many people experience in their daily lives, and they wouldn’t be crammed together like many people are on tubes or buses.... But even if it’s accepted that there is likely to be more transmission on a cruise ship than in typical daily life, only two weeks elapsed between the infected passenger alighting and the ship being quarantined. Assuming quarantining people to cabins had a significant, though not total, suppression of transmission (much like in our lockdown), the R0 would need to huge (akin to measles and well in excess of even upper estimates) to be confident that all, or nearly all, passengers had been infected.

I would like to see more investigation and data into why certain geograhpical regions / groups of countries do better than others, whatever they do

What is different about Scandinavian countries, compared to N&W European ones ?

Sweden does much worse than its Scandi neighbours,
but Sweden's death rate is still is comparable to the better countries with lockdowns in N&W Europe
So they didn't need lockdown, but e.g. Italy, France, Spain, NL, Belgium, UK did

Scandi countries that did lock down have v v low death rates, better than even Germany

Unfortunately ^the UK behaves like the Italy group, following those curves
and not the Scandi group^

Are there habits / diet / whatever that the UK - and all the other N&W countries - can change, to get the Scandi levels ?

I have real doubts about using the Diamond Princess and Theodore Roosevelt as examples of the spread of Coronavirus except in relation to their own unique circumstances.

I definitely would not base any lockdown exit strategy on these limited idiosyncratic examples but I am hoping researchers do currently work on assessing percentages of population who may have innate resistance to C19.

I am also hoping there is work going on to understand the reasons for London peak of hospital deaths on April 4th. If the R0 reduction was indeed self-triggered prior to strict lockdown, it appears that some of social distancing measures (e.g. work from home) are much more important than others (e.g. school closures).

Anyone else concerned that the government focus on hospital deaths as the key figure hides the stats on effectiveness of hospital treatment?

My concerns are:

1. It creates an incentive at a system level to minimise the numbers admitted to hospital. This may impact survival rates significantly, especially if early admission eg for oxygen helps survival

1. It completely hides the figure of death rate as a percentage of overall cases in the U.K. Without wider testing we are simply estimating cases based on death rate. So how on earth can we even tell if our deaths per cases are higher/lower/the same as eg Germany? How can we learn and adjust things like when admitted, etc etc in order to improve our figure if we have no idea what it is?

1. Given the strategy is to not exceed NHS capacity, there is a drive/pressure by some ministers to ‘run this hot’ - ie lift lockdown measures to the extent the NHS can cope. With no way of knowing how many more are dying in the U.K. than elsewhere, what are the break measures to counteract this and ensure anyone who might need earlier treatment is afforded precious nhs space?

A point made in this morning's More or Less programme is that if the peak of deaths in the UK was 8 April, this suggests the peak in infections was 3 weeks before that, 18 March, full lockdown didn't happen until 23 March, so was it the pre-lockdown measures that led to deaths starting to fall?

How effective is full lockdown over the kind of social distancing measures that people are willing to adopt themselves? I guess that's the question being played out in Sweden and what's at issue in the US.

I'm still a fan that harsh measures were applied, albeit too late, than had we gone for incrementally strengthening measures from a low level. Better to row-back from too harsh in my view.

Germany includes deaths at care homes and any institutions like prisons etc
because these are automatically sent electronically to RKI (the German public health authority)

but not deaths of people at home,
who didn't bother / want to get tested, registered and monitored
because they are not on any official system

Deaths at home in any country are probably something that one can only add to the count several weeks later,
sometimes having first to somehow find the corpses of those living alone and socially isolated

No excuse though for excluding care homes and other institutions ,
even if this needs to be done say 2 weeks later, once the data is there

I hope at least the UK govt advisers have access to this data and collate it all to form their recommendations

Anyone else concerned that the government focus on hospital deaths as the key figure hides the stats on effectiveness of hospital treatment?

I am afraid the info on hospital deaths is the only data point available on a daily basis. We can also use these numbers to compare our trends with Italy and Spain (who are ahead of us on the curve and also only report hospital deaths).

The comprehensive death numbers are available in weekly ONS reports that seem to be now getting quite a bit of attention and press coverage. It would have been very useful to have this data closer to real time but, unfortunately, while ONS may be reliable, it is also slow and not exactly efficient.

That FT curve from a while back showed a clear relationship between early lockdown and better outcomes

It is likely that harsh measures done early enough are the most "efficient:"

the curves stay at a lower level than otherwise

... and hence they can be relaxed after a shorter time,
to be replaced by lesser measures that allow 90% of the economy to resume

e.g. masks outside, keeping the ban on adult crowds > 50 even outdoors

Rather like for an individual who doesn't get treated early enough:
the disease takes much longer to go
and the person may also be left with much more serious longterm effects

Yes, agree there is no reason not to have more timely reporting for care homes.

France ended up reporting those numbers in the daily stats as they started getting higher and comparable to hospital deaths.

More or Less on Radio 4 was very interesting.

It was reassuring about the risks to NHS workers- I'd been asking myself the same question yesterday.

Bus drivers are overly represented in the death rates. So that's something that needs serious thought.

agree there is no reason not to have more timely reporting for care homes

Deaths need certifying, may need to go to the coroner, may need follow up testing and double checking. Deaths at home also take a long time to certify. Anyone who dies outside of a hospital environment can take weeks and months to get answers on "unclear" deaths. That's just the way it is. Even hospital deaths can wait weeks to be "certified". This isn't because they haven't died, of course they have, but they won't be included in the stats if there are further examinations into death to be done.

Bigchoc - there are no perfect options here. If we try and get coronavirus deaths to zero, that would only be at the expense of enormous suffering. It would require a full lockdown until a vaccine, which would lead to enormous unemployment, young people with huge educational gaps, and probably starvation of millions in developing countries as rich countries stop buying anything that is not strictly essential.

If you think that that is too much, then the question is what is proportionate. It is simply fact that people under 40 with no underlying conditions face almost zero risk from this virus. I’m over 40 so at a more modest risk, but I know enough about statistics to be relaxed.

We have evidence from Sweden that their case rate is not exponential despite bars, restaurants and gyms being open. I know that you think Scandinavia is different, but I am still at a loss as to why their bars, restaurants, hairdressers and supermarkets differ from ours.

We know bus drivers and HCPs without PPE are at risk - but has anyone investigated others?
For example I am hearing most days about a teacher dying. I know several teachers who have been quite poorly, but as far as I know noone is researching this. I know primary children are less likely to catch the virus, but what about secondary?

crunchycarrot are you still doing the chart that compares us to Italy? I'm wondering how we're doing on that front.

Yes, but not up to date on the Italian front. Also not sure whether the Italian figures are hospital only, like ours? Plus we have far more deaths than are being included, due to care homes and at home deaths.

On the next thread, I'd like to include in the OP a simple criteria of which countries DO include all deaths vs just hospital deaths so we can see when we are comparing apples to oranges.
Also perhaps some metric of testing, eg Germany tests X thousand weekly vs UK tests a tenth of that.

I don't have that data but if anyone has a link to a handy compilation I'll include it as a link in the next OP.

A point made in this morning's More or Less programme is that if the peak of deaths in the UK was 8 April, this suggests the peak in infections was 3 weeks before that, 18 March, full lockdown didn't happen until 23 March, so was it the pre-lockdown measures that led to deaths starting to fall?

I think "3 weeks" is an approximation

I think it’s more the case that people have died faster here — many of the vulnerable have not been put on ventilation. Just different approaches to end of life care.

I think "3 weeks" is an approximation

Of cousre. The point is that if any kind of model doesn't work, then you need to find out what's wrong with the model. If lockdown, rather than previous measures, did lead to the reduction in deaths, that means people die relatively quickly in the UK. Why would that be? Would it be that the vulnerable aren't put on ventilation as EmMac7 suggests?

I think that many units are trying to avoid ventilation in general, not because of rationing but because they are associated with poorer outcomes.

Normally deaths are a lagging indicator as even if the average time from infection to death is 3 weeks, more people are likely to take longer than shorter. That is, you would expect fewer people to die suddenly within a week but for more to survive for longer - e.g. by being kept on a ventilator. This can be seen in the death charts - the rise is sharp but then they plateau rather than having an equally sharp decrease.

It certainly is worth considering the implications of an 8th April peak and investigating it, rather than dismissing it. There seem to be a lot of people here who dismiss any indications of anything other than ‘hide and wait for a vaccine’ with emotive terms rather than rational analysis.

My office went to working from home from 17th March, after Boris’s speech on the 16th. We are in central London so consistent with social distancing for adults. My children stayed at school - including travelling by tube - until the 20th. Tube travel was much lower then but not as low as now.

It is certainly interesting if that already had an impact as it could suggest that we could go back to the March 17th world, before full lockdown, without exponential growth in cases.

It is certainly interesting if that already had an impact as it could suggest that we could go back to the March 17th world, before full lockdown, without exponential growth in cases.

Yes I agree. Looking at the government data on transport use change, it fell very sharply in the week before the lockdown. I would imagine that this, along with other data, is being interrogated to see which changes in behaviour (especially voluntary changes) and lockdown measures have had the greatest effects on infection rates. This should point the way to the safest exit strategy.

That's what I think this thread is about. What are the real numbers and what do they tell us about what's actually going on rather than the narrative we're told for whatever reason, ie political etc.

I am in London
Public transport usage is linked to school and work commuting
One of my DC’s schools in London closed midway through the week before lockdown.
People in financial services, the law etc already wfh on an occasional or regular basis so a switch to total wfh with the knock on effect on commuting was easy.