Daily numbers, graphs, analysis thread 3

[Barracker]

Welcome to thread 3 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 is a key piece of data but we don’t have it. I seem to believe that the Health Secretary said they are trying to find another method to report recovered.

Our data is very shoddy but I think we have peaked or are peaking. If you smooth the data, there is very little increase in new cases over the last week or so.

I have been following this thread as there are many great contributors helping me make sense of the trend. I am puzzled by the fixation on daily deaths. They are definitively the end of the line. Given our woeful lack of testing, surely the first indicator of a turnaround is the hospital admission rates.

I have just finished the Reuter’s analysis of why Johnson was so slow in getting to grips with the huge danger Britain faced.

www.reuters.com/article/us-health-coronavirus-britain-path-speci-idUSKBN21P1VF

Just a personal anecdote.
We live in Luxembourg and my partner was booked for a trip to London for four days on Monday 17th February.
The week before my partner was extremely nervous about travelling to London because of the world wide Corona Virus epidemic. She would have cancelled the trip but the flights and Hotel were pre paid and I persuaded her to go!
I bought from our local pharmacy face masks (they had to be ordered as they were out of stock) which arrived on Friday 14th February.
She travelled and had a good time despite her fears of the Virus and returned safely.
She was right and I was wrong. However I am not the Prime Minister.

I have been looking in Worldometer and it looks like Sweden might have plateaued in terms of new confirmed cases already. Without any lockdown. Am I right in this? Or am I misreading?

It depends on how many tests they are doing. Confirmed cases don’t tell you much on their own. Their deaths are equivalent to 6-700 per day in the UK.

Plus today Sweden reported more CV cases than any day so far... That isn’t enough on its own to indicate things are getting worse (depends on whether scope of testing has changed, how they are reported etc.) but it similarly doesn’t indicate things are stabilising.

Our data is very shoddy but I think we have peaked or are peaking. If you smooth the data, there is very little increase in new cases over the last week or so.

I don't think we are peaking yet - statistically we have at least 5 days to go (probably 9). However, if the government's initial "wash your hands and stay at home when ill" was at all effective we would be seeing the results of that by now, in a slowing of growth. I think that is what we are seeing/will see. A slower growth (than other countries) from this point before we reach our peak.

Oakmaiden yes that sounds a fair point

"The whole point of lockdown I s to get the transmission rate below 1, so ideally the peak new infections actually happen right when lockdown is called because they start to drop at that point, and what we see delayed in the graphs is peak diagnoses/ICU admissions/deaths. And that can’t be later than 3-ish weeks after lockdown unless lockdown isn’t working."

Peak deaths 4 weeks after lockdown is quite possible.
Or longer

It was about that for Italy
Will be about that for Germany - Lockdown on 16 March and peak expected just after Easter

I think lockdown timing may be different depending how far along the infection has progressed in different areas - e.g. whether early early on or too late, as in Italy -

how much people are still moving around,
which includes essential workers and also all the others who can't WFH but have to work to get paid

and maybe how long seriously ill people can be kept alive

Someone can be infectious for up to 14 days, possibly without realising
The person they infect can take 14 days to die - or be kept alive longer on ICU-

Sweden had 96 new cases; their neighbour Norway only 12 cases
Their rates are definitely higher than their neighbours and the trend looks more like increasing still

Although today's deaths were lower than yesterday's, they may have high Tuesday figures for similar reasons to ours - the weekend effect

We generally need a week to be sure of a curve flattening
Too soon to say about Sweden

Scotland has a new report out

www.nrscotland.gov.uk/covid19stats

The key points which journalists are distorting are that last week had UNDER reporting of deaths due to a change in reporting methods, so this week's total increase in ALL deaths of 60% is partly deferred deaths from last week.

However despite that, there is a total increase in deaths of 27% for both weeks. That is around 550 extra deaths compared to the 62 last week and 282 this week from covid-19, so presumably some unrecorded covid-19 deaths as well as deaths from other causes not getting medical attention etc.

Essentially a 10% increase in ALL deaths over and above the covid-19 deaths is implied, but presumably more than 10% for this week (to 5 April registration) and less for the previous.

Anyone who thinks that we are somehow not going to end up with an increase in total deaths for 2020 is living in fantasy land.

That said if we end up with a mere 20,000 extra deaths for the UK this really is no big deal as a number, it's more about what it could have been, and whether we can keep things that low, which seems unlikely....

WorriedMutha
I am not sure there is a clear measure for hospital admission rates. The testing data is getting clearer but confirmed cases may also include NHS staff who do not need hospitalisation. So confirmed cases could climb sharply as testing ramps up without reflecting a worsening of the epidemic.

Oakmaiden,

When I say peaked I mean new cases, not deaths, and it seems to me that we are there or thereabouts.

What statistics make you believe 5-9 days from now?

I'm not sure how much to believe in the cases peak for the UK, given the limited testing

Sadly, deaths looks the reliable metric to follow

I can't see the peak yet:

April 6
3802 new cases and 439 new deaths
April 7
3634 new cases and 786 new deaths
April 8
5491 new cases and 938 new deaths

@Derbygerbil it seems that Sweden have a lower death rate than ours at the moment, so would assume testing at a similar rate? Maybe this is a wrong assumption. I just wonder if that level of social distancing us actually enough to bring the infection rate to below 1? It their rates are similar, and remain similar to other European countries what would be the conclusions? I was looking again at the graphs, it does look like its flattening for new infections.

Bigchoc some of the deaths reported today will actually have been deaths occurring on Friday, Saturday and Sunday. There is a lag in reporting over the weekends.

I know of the reporting delay, but the stats we have still don't show a flattening
We don't have any others to go on

It would be better to compare Sweden to its Nordic neighbours, with whom they have far more in common
rather than to the UK, Italy, France

and compared to e.g. Norway, Sweden has been doing much worse

However, even their "much worse" may still be acceptable, if the curve flattens soon,
but we're not seeing that yet:

April 5
387 new cases and 28 new deaths
April 6
376 new cases and 76 new deaths
April 7
487 new cases and 114 new deaths
April 8
726 new cases and 96 new deaths

Quick question...there's a thread where Crunchy's tables are being called "arbitrary" and meaningless because they start at the 10th death.
I remember there is a reason for that, and it's the way experts track comparisons as well. Can someone remind me?

All countries will have a reporting lag, but we take the peak using the stats for that "nominal" day,

because it is too complicated otherwise to say it is really X0 for that day + X1 for the day before + X2 for the day before that

I agree bigchoc also the community deaths are only being counted weekly by ONS we don’t have a true posture of the infection rate or the death rate. Going purely on confirmed cases (hospital admissions) of 60,000 and deaths of over 6,000 it shows a mortality rate of over 11%. However it’s likely to be lower than that because of mild/moderate cases that aren’t tested. I doubt we will ever know the true extent of the spread.

I also think that we aren’t near peak yet but the massive jumps after the weekends really aren’t helpful when looking at trends.

Picture not posture!

I read on another thread that the numbers read out at the briefing aren't the actual figures as they don't include everyone that died. Is that true? Can't be can it?

The UK's growth rate doesn't seem to be responding to the lockdown, unlike Italy's

The Reuters article is an interesting read Gfplux.

I'm particulary taken by the following paragraphs

  <span class="italic">According to one senior Conservative Party politician, who was officially briefed as the crisis unfolded, the close involvement in the response to the coronavirus of the same scientific advisers and civil servants who drew up the flu plan may have created a “cognitive bias.”</span>

  <span class="italic">“We had in our minds that COVID-19 was a nasty flu and needed to be treated as such,” he said. “The implication was it was a disease that could not be stopped and that it was ultimately not that deadly.”</span>

I was involved in planning for SARS and Swine Flu and limited revisions for MERS and my suspicion is that Covid 19 was seen as a "nasty disease that happened to chinese people" and it was assumed that like the previous episodes and ebola, the UK and most of the western world wouldn't be seriously affected. The trouble is, this time, the wolf has turned out to be real.