Daily numbers, graphs, analysis thread 10

[BigChocFrenzy]

Welcome to thread 10 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

We welcome factual, data driven, and civil discussions from all contributors 💐

I disagree that the daily tests are meaningless now that we test a lot of people, they are the fastest indicator we have. They were pretty meaningless at the start.

Unless the time lag between tests and reporting changes it doesnt affect the trend.

That's interesting @sunshine. So do you think that the data on hospital admissions is a better indicator than positive tests (albeit a week or so behind)?

The ONS released this week was very encouraging, though as others have pointed out its small sample will mean it's not great at picking up small changes, plus there's a bit of a delay on it.

It's annoying that they don't report the numbers of tests carried out each day. They must know this because they know the number of positive ones so could just as easily collect the negative results alongside. That would make it much easier to see whether rates are going up, or just testing getting faster (anecdotally, they have got a bit faster in the last week or so)

Test data:
Number of tests carried out per day
Number of self tests as % of the above
Number positive
Number negative
Number inconclusive
Number whose result has changed since last test

without that all test data is pissing in the wind

if HCP and Care staff are not being tested weekly
and a 0.1% sample (56,000) are not being tested daily

the data is anecdote garbage
not science

I mean trying to read anything into the daily data on its own is fairly meaningless. The trends have meaning, daily figures don’t tell us very much.

@alreadytaken

I disagree that the daily tests are meaningless now that we test a lot of people, they are the fastest indicator we have. They were pretty meaningless at the start.

Unless the time lag between tests and reporting changes it doesnt affect the trend.

Currently they are meaningless, because we don't know how many people are being tested, or when those tests were taken. That 1,514 bears little or no relation to the number of tests supposedly "conducted" yesterday.

If the amount of people we are testing is increasing (likely), then an increasing number of positive tests isn't a concern assuming they aren't increasing in percentage. But we don't know.

The time lag between testing and result is, I think, supposed to be getting shorter, so it's more likely that the tests reflect a more recent position than they did, but they come from so many different sources now they are likely to vary widely.

The best actual data we have at the moment is hospital admissions by region, which we should get an update on tomorrow (though that has regular weekly dips over the weekend too).

we dont know the numbers tested because the government wont tell us and they no longer give hospital admissions figures at weekends. I dont pay any attention to the number of tests taken - but I'm sure someone can do a trend figure.

@Firefliess

They do announce the number of tests per day but they have stopped announcing the number of people tested (some people are tested more than once) which would be much more useful information!

Got curious, so here we go - UK positives as a percentage of tests conducted for P1 and P2, and as a % of people tested for P1, and seven day rolling averages for those three. For what it's worth....

P1, as a reminder, is swab testing in Public Health England (PHE) labs and NHS hospitals for those with a clinical need, and health and care workers, and makes up about a third of current testing.

What this seems to show is overall a trend down, but some flattening out over the past week. Spikes are probably down to a backlog of testing results arriving on that day, resulting in an increase in positives against a normal number of tests.

@sunshine What they tell us is not the number of tests actually carried out though, it includes the tests that were posted out each day. They'll turn up as positive tests in some later day's figure (if at all), but you can't tell which day (which I fear makes those nice charts someone has just produced not quite work either)

It would be much easier to see if numbers are falling or not if they gave us the number positive, the number negative and the number inconclusive/failed for each day. Then you could see if increased positives was due to increased testing. I can't believe they're not collecting that data - they're just not sharing it because they prefer to share the larger number of "tests" which includes those posted out.

@Firefliess that's why I said "for what it's worth" !

The number of tests reported as conducted bears little relation each day to the number of positives. I was just curious to see whether there was anything that could be gleaned from a trend over the past 6 weeks. TBH, it's not as all over the place as I thought it might be.

We do have some data on positives for P1, per person based on sample date, and it's declining rapidly, but it's such as small number of tests it's not really that interesting.

@patricia I'd guess that the pillar 1 declining is due to more routine testing of NHS staff (who have a much lower % positive than patients, obviously) My friend who's a nurse has been tested a bit recently

[quote Firefliess]@patricia I'd guess that the pillar 1 declining is due to more routine testing of NHS staff (who have a much lower % positive than patients, obviously) My friend who's a nurse has been tested a bit recently[/quote]
Yes, I should think so. Pillar 2 - testing in the community - is really what we need more detail on but just haven't got.

In a study of COVID-19 infection clusters in Japan, approx 50% of non-healthcare clusters were seeded by 20-39 year olds mainly asymptomatic or pre symptomatic. Clusters were often associated with heavy breathing and or close proximity, such as singing karaoke, talking at bars, or exercising at gyms.

wwwnc.cdc.gov/eid/article/26/9/20-2272_article

Long time lurker here. I find these threads really fascinating and I'm in awe of the level of statistical analysis from many of you.

Just wanted to mention that among some of the people testing positive, might it be a false positive, as in someone who has had the virus and might still be shedding it (hence the positive test) but is no longer actively infectious? Probably quite a small number, but I know some people report symptoms fluctuating over long periods who are very unlikely to still be infectious.
I've just been requested to have a test through the Zoe app. I've had various very mild symptoms on and off since late March. No idea if it's Covid or not but if it is I must have picked it up weeks ago.

There is a % of false positives in swab tests
and a probably larger % of false negatives - swab not pushed far enough in
and I've also read because virus particles have moved down into lungs

Anyway, whatever the bug, I hope it finally goes and leaves you in peace, fadingfast

It's interesting that spread seems to be increased where deep / heavy breathing is involved.

I hope this is taken into account when unlocking measures are made. I'm thinking that things like singing, gyms, sports classes, school PE, nightclubs, dance classes etc would need to be considered.

I do despair a bit in that I'm constantly trying to keep tabs on the data but unless you have all day to look at it I miss stuff - e.g. I saw that the daily data tracker Local Authority case numbers etc no way near added up to the total cases announced but I didn't realise it was because it's only Pillar 1 positive cases (hospitals) not Pillar 2 (wider community... I think? I'm a little confused in the same way as the previous poster upthread).

With so many data sources, caveats, missing data and background knowledge needed about all of this it's so hard for me to build up any sort of picture.
And inaccuracy of tests as well!
Sorry, just a little frustrated rant.

What'snext Japan is interesting as they wear face masks and have hand sanitiser etc so much everywhere anyway day to day that everyone automatically upped their usual routine the Minuit this came off the scene.

We were speaking to our friends in japan yesterday about this.

That study is notable because in japan, those are the times you wouldn't be wearing a mask.

(I also have mused about the Vit K link as Natto is very high in this. No idea if it's a really common food though.)

I also have mused about the Vit K link as Natto is very high in this. No idea if it's a really common food though

I believe it is very much a feature of the Japanese diet, Neurotrash. I tried it once in Japan but retired defeated - the appearance was the first hurdle, the texture a second and the smell another (I should add that Japanese cuisine is generally wonderful, and I'm sure there are things we eat that would be unpalatable to Japanese people).

Can I add to the wish list for data?
As well as test numbers (positive, negative, total tested), dates of testing and where; - numbers for traced and contacted and numbers in isolation.

and clear numbers of how many actual people tested

• we don't know the average numbers of tests / person, which may have changed over the weeks

Gyms have been open in Germany since at least 16 May, with some limited social distancing

• eveyone brings their own mat for classes, reduced numbers in classes, machines & benches spread out Also sanitisers everywhere and nightly deep cleans

So far, no outbreaks around gyms,
although there have been in a few care homes & geriatric hospitals, meat plants, package depots, churches - singing !

What is noticeable is that in one state which chose to increase the number of people that couold meet from 10 to 50
(German states have a lot of autonomy)
there have been outbreaks after large private family parties, with some family members hospitalised

btw, I've been going to my gym ever since the 16 May reopening

I'm nearly 64, but with the low numbers of cases & deaths in total, I decided it was a sufficiently low risk for the gain in pleasure & fitness

I'm also going to start eating on Fridays at outside tables at restaurants

Has there been any outbreaks in shops or supermarkets? I am frequently told that this is high risk but I have not known or heard of any outbreaks amongst shop staff.