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.

Thanks for the volcano, Barracker
I always look forward to that

That 31 March glitch bothers me too, because we depend on data being reasonably correct
< and I'm a hardcore Aspie, so my brain won't let go >

. WHY hasn't it been either corrected, or explained in footnotes if it is legit ?
e.g. a switchover to another reporting / dating system.
Many in PHE - and ONS too - must have noticed this anomaly. Even a footnote to say the numbers will be updated later

. What lesser glitches are present, not so obvious, that we haven't spotted ?

. Will it happen again, e.g. on 30 April or 31 May, if there is a bug with EOM or EOM with31 ?
Or 30 June if the issue is end of quarter-year

Thank you for your updated graph @Barracker - it’s a lot of work and much appreciated x

Move Some people who had winter bugs incl flu may indeed hope they had COVID

• and hence probable immunity - immunity, but as sofa said, we'd have seen early COVID in the deaths 2 weeks+ later

This winter, flu deaths etc were low, so that deaths were in fact lower than average over winter
A sudden jump in deaths from undiagnosed COVID - respiratory disease, maybe CVD - would have been obvious

Unnoticed COVID would mean it was a very mild form, then mutated very quickly to become much more virulent,
which isn't what happened in China and isn't what such viruses normally do

Scientists say that the different COVID mutations don't vary as significantly in their makeup or effects compared to say flu,
which is one reason they are optimistic about a vaccine

Barracker MPs can still ask questions of the civil service / ministers
So maybe you could ask your MP to enquire about this 31 March data anomoly ?
(I don't have an MP, or I would)

There were 2 cases recorded by test on 31/01/20 and the first death recorded as Covid was 10 March. I'm not sure when they started testing as the graphs seem to conflict. I'm also not convinced that Covid would have been given as Cause of death before March in a lot of cases. As I have said previously, given that London is an International flight hub, out of the thousands that left Wuhan before lockdown, it is very probably that a good number were on flights that passed through Heathrow.
coronavirus.data.gov.uk/#regions

Cluster of COVID-19 in northern France: A retrospective closed cohort study

This is a COVID epicentre, with a high % of infected, 25.9%
BUT
only 17% of these had no symptoms

This is lower than some other studies, but maybe a significant % people with partial innate resistance spend several weeks in the "Pre-symptomatic" stage before having symptoms ?

https://www.medrxiv.org/content/10.1101/2020.04.18.20071134v1

The proportion of infected individuals who had no symptoms during the study period was 17.0% (95% CI = 11.2-23.4).

The proportion of donors with anti-SARS-CoV-2 antibodies in two nearby blood banks of the Oise department was 3.0% (95% CI = 1.1-6.4).

Interpretation:
The relatively low IAR observed in an area where SARS-CoV-2 actively circulated weeks before confinement measures indicates that establishing herd immunity will take time,
and that lifting these measures in France will be long and complex.

The ONS deaths from the start of the year are lower than the average over previous years
and show no indication of extra deaths that could have been COVID

The curve only rises sharply at a time consistent with the curve of COVID deaths

Here's a scatter graph showing Index of Multiple Deprivation vs. death rate for London boroughs. There is some correlation, but e.g. Bexley is slightly less deprived than Harrow but has half the death rate.

If the peak of deaths was on April 8th, does this mean that the peak of infection rate happened during mid-March?

shoots It's difficult to see significant correlations with deaths

What about travel / transport links ?
but I don't know parameters to plot that, even in London - distance to tube stations doesn't look feasible

Mum The timing would fit with lockdown and the week or two earlier when lesser social distancing, hand-washing etc started

If the peak of deaths was on April 8th, does this mean that the peak of infection rate happened during mid-March?

Just before lockdown and the schools closing?

Does that mean that it’ll happen again if reopen as before?

WHY hasn't it been either corrected, or explained in footnotes if it is legit ?
e.g. a switchover to another reporting / dating system.

Hi all - been mostly lurking but I have appreciated the information shared here very much.

Re the above- I do remember a point when they announced they'd changed the cutoff times and one day's data had either more or fewer cases as a result. Can't remember if it was deaths or cases or if data was backdated. What am I thinking of there? I assume that wasn't the 31 Mar thing?

There was something in the reporting at or around 31st Mar as that was when NHS England started issuing date of death breakdowns. What is the source for reporting date/occurred date splits of figures going back earlier?

@BigChocFrenzy

I read somewhere(forget) that there was a low infectious disease transmission rate at the end of March. It dropped significantly.
Which means that hand-washing and less restrictions were enough?

And then there’s this professor who is saying that virus’s peak with the same results with or without draconian measures. Could he be right?

www.telegraph.co.uk/news/2020/04/23/coronavirus-dies-within-70-days-no-matter-tackle-claims-professor/

“Across the globe, debate is raging about the best way to tackle the spread of coronavirus, with countries adopting radically different approaches in the fight against the disease.

But an Israeli professor claims all efforts will lead to the same result, because the disease is self-limiting and largely vanishes after 70 days – with or without any interventions.

Professor Isaac Ben-Israel, the head of the security studies programme at Tel Aviv University and the chairman of the National Council for Research and Development, claims his analysis proves that Covid-19 peaks at 40 days before rapidly declining.

Major General Ben-Israel, who was also head of the analysis and assessment division of the Israeli Air Force Intelligence Directorate and former chief Cybernetics adviser to Israeli prime minister Benjamin Netanyahu, claims shutting down major economies is having devastating consequences for little gain.

Based on the new data, the team at Tel Aviv University has called on the Israeli government to start a phased lifting of its restrictions while keeping social distancing in place. "Given that the evidence reveals that the corona disease declines even without a complete lockdown, it is recommendable to reverse the current policy and remove the lockdown," the researchers said.”

redlock It would probably happen again if all measures were relaxed at once
i.e. sudden switch from lockdown to "normal"

The strategy in other European countries is to relax measures very slowly in steps, guided by scientific advice at every stage
and be ready to slam on the brakes if exponential growth starts off again.

It is described in Germany (where I live) as walking a tightrope
with a worst case of 1 million German deaths if we fall off

Major public / cultural events like Oktoberfest have already been cancelled
and we have been warned that some social distancing will likely remain until there is are reliable treatements or a vaccine.

However, we can't stay in lockdown until then;
people have to pay bills and kids must go to school

The strategy here is to gradually get 90% of the economy going, but not the other 10% which involves large groups of people - mass entertainment, flights
That 90% still means adapting with some social distancing as far as possible

Germany has a huge program of mass testing, contact tracing and isolating the infected
which is hoped will both reduce the chance of exponential growth stating again
and give early warning if it does,
so Merkel can slam on the emergency brake

mumlove5 That Israeli professor seems to be a mathematician, not an epidemiologist

Epidemiologists say he just seems to be doing math curve-fitting & extraplation and has not put into any model how epidemics actually behave

However, phased lifting of restrictions is what every government is planning, so that recommendation is nothing new

We obviously can't stay 2 years in lockdown,
so each country will have an "acceptable" death rate and sub-1 R0, at which point they will trisk relaxing measures in stages

If the virus was circulating in Jan, making a number of people ill but not markedly so (but enough to perhaps stay off work for a few days), not being tested for and causing relatively few deaths, with those attributed to other causes against a lower than usual baseline (perhaps among people expected to die soon anyway) - it wouldn’t show up in the statistics anyway. We don’t know how long the lead in curve prior to testing might have been, before it started rising steeply.

From John Burn-Murdoch@jburnmurdoch (FT)

(late !) Fri 24 April update of coronavirus trajectories

Daily and cumulative deaths
Week-on-week hospital death

Subnational regions & city deaths, daily & cumulative:

@BigChocFrenzy

But could the Israeli professor be correct?

This probably sounds insensitive but I don’t mean it that way... the original Imperial Model says that with mitigation measures the UK will see 250K deaths. Obviously, we’re no where near that. I’m just not seeing the numbers supporting draconian lockdowns.

If peak infection transmission rate was indeed before the lockdown, some social distancing measures were enough.

This could be just the natural progression of viruses. They peak and go down no matter what.

Plus, flu season ends mid-April again leaning towards natural progression.

@Redlocks30, they won't open the schools or change anything until the R number is where they want it to be. This is to prevent a second peak. Presumably, before the lockdown, say the last week the schools were open, the R number was high, hence the lockdown. So, there might be a rise when they soften the lockdown/measures, but it shouldn't be massive. Getting the R number down to where they deem it safe is key to this.

You can see the Respiratory Infection chart in this link.

www.cebm.net/covid-19/what-does-rcgp-surveillance-tell-us-about-covid-19-in-the-community/

Many things are possible, but we don't have evidence they actually happened

• and may never do

Even once we get results of national antibody studies, it may still be difficult to either prove or rule out any theory

If COVID has an R0 of up to 5.7, as in that recent CDC paper, then it may have spread much more than expected.
Or the R0 is 2.0 and it hasn't

That French paper of only 17% infected without symptoms after a few weeks
... is concerning
We need more data than 1 study though, which could be anomalous / flawed

The data highlight that these initial distancing measures, together with the seasonal effect, reduced transmission of URTIs by 9 per 10,000 (44%) in the week (from 20.4 to 11.4 per 10,000 consultations). The following week (22nd March), when the lockdown was introduced, rates of URTIs further decreased by 3.3 per 10,000 consultations (29%).

COVID cases were first detected in week 10 (8th March). The initial rates of 0.23 per 10,000 now stand at 6.43 per 10,000 by the 19th of April which is little change from last week’s rate of 6.34.

mumlove5 The Imperial estimate of ¼ million was for much lesser measures than lockdown

... but we did lockdown, so we didn't get that ¼ million

They also predicted ½ million if we did nothing at all

So we'd expect to see a little difference between doing nothing and doing something, but not enough

The peak flattened very slowly, so that is not evidence that mild measures were enough,
just that they had some effect

The only comparison between lockdown / not is in Scandinavia, where all Scandi countries have much lower infection rates and deaths,
for reasons the scientists have yet to decide / publish

BUT
Sweden - without a lockdown - has a death rate several times those of its neighbours who did lockdown

The UK is on the Italy curves, not the Scandi curves,
so multiples of Italy, with exponential growth, looks reckless