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 📈📶👍

[quote whatsnext2]Interesting comparison of Covid against flu. Basically similar to pandemic (as opposed to endemic) flu apart from age distribution.

threadreaderapp.com/thread/1284513419454971905.html[/quote]
I disagree with those comparisons.

The major reason Western countries were so unprepared for COVID is that they prepared for a flu pandemic and COVID has very different characteristics to the flu:

• IFR for annual flu - not the 1918 or 1957 pandemics before our time - is 0.044%
  Most virologists & public health officials estimate COVID has about 10 x the IFR of seasonal flu

• COVID is infectious for up to a week or more before symptoms appear
  Hence the SD, masks etc which would not be so suitable for seasonal flu

• COVID places much higher demands on hospitals - about 20% of confirmed cases need hospitalisation
  . This means that a sudden large number of infections can overwhelm even a goo standard health service, as in N Italy
  . It also requires far larger stocks of PPE, which is why the flu pandemic stocks of PPE in Western countries were totally inadequate

• We still have no date for a COVID vaccine; seasonal flu has an annual vaccine
  The plans for a flu pandemic were to hold out for 6 months until existing flu vaccines could be tweaked and rolled out

• Flu R0 is about 2, whereas COVID R0 is usually estimated at over 3 - that's without any measures;
  In fact, the CDC has estimated 5.7

• COVID has a much lower K (dispersion factor) than seasonal flu or 1918 flu,
  about 0.1 compared to 1.0
  Hence about 10% of cases are responsible for 80-90% of further spread
  i.e. superspreaders, who can infect 10 or even hundreds of people, while most people infect 0 or 1 person.
  Again, this requires different SD measures to flu which spreads much more uniformly

Positive news that should help councils control local outbreaks
Announcement next week, possibly on Monday

www.theguardian.com/politics/2020/jul/18/matt-hancock-in-new-u-turn-on-coronavirus-testing-data

The health secretary, Matt Hancock, has bowed to pressure from councils,
which demanded full access to the names and data of people in their areas who tested positive for Covid-19,
and those with whom they have been in contact,

Rollout of easier & quicker antibody tests:

https://www.theguardian.com/uk-news/2020/jul/18/uk-plans-millions-of-coronavirus-antibody-tests-after-trial-success-report

The UK government is planning to distribute millions of free coronavirus antibody tests after successful secret trials,
according to reports.

The finger-prick tests, which can tell within 20 minutes if a person has ever been exposed to the coronavirus, were found to be 98.6% accurate
in human trials held in June
...
“While these tests will help us better understand how coronavirus is spreading across the country,
we do not yet know whether antibodies indicate immunity from reinfection or transmission,”

Can somebody please explain something to me?
I have heard the criticisms that the UK are recording deaths as Covid which may not be due to previous positive tests which suggests over-reporting but I have also seen that excess deaths are far higher than recorded deaths. Can these two things be reconciled? How can they both be true?

Do we have any reliable data on the percentage of false negative Covid-19 tests? (Antigen tests, not antibody)

@CalmYoBadSelf

Can somebody please explain something to me? I have heard the criticisms that the UK are recording deaths as Covid which may not be due to previous positive tests which suggests over-reporting but I have also seen that excess deaths are far higher than recorded deaths. Can these two things be reconciled? How can they both be true?

Both happened, but the under-counting was far greater

While there were a lot of COVID deaths, March - end May, COVID deaths were under-reported

• we know this from ONS reports and the 20,000 gap difference in excess deaths & COVID ones

(A few deaths may still be missed e.g. of the elderly not tested)

However, PHE have been counting all deaths outside the 28 day limit for a positive test and some of those will not be COVID.
Their initial figures indicate they over-counted by a total of 2,200 deaths

Excess deaths have been in the normal range for weeks now, with COVID deaths averaging under 100,
so this over-counting matters for the daily PHE figures

The ONS data, including of excess deaths, remains the gold standard and will not be affected by this PHE decision

@calm Early on in the outbreak we were not testing enough people and were underrecording Covid deaths as a result, particularly in care homes. We know this because the excess deaths were about 10,000 more then the Covid deaths (and all the data on the age profile of the excess deaths, etc very much suggests that they were predominately Covid deaths not any indirect effects of lockdown).

Later, we sorted that problem out and started testing more people so there are no longer excess deaths which aren't being recorded as Covid. However in an effort not to miss anyone we started recording everyone who has ever tested positive for Covid as a Covid death, even if they had it months ago, recovered fully and died of something else. This is starting to affect the daily death figures more as the genuine numbers fall and the pool of people who had it months ago and might just happen to die of something else grows.

Overall we have under reported Covid deaths, but in the past few weeks we've probably been over reporting

CMMID: Reconstructing the global dynamics of unreported COVID-19 cases and infections

https://cmmid.github.io/topics/covid19/Under-Reporting.html

Conclusions:
We found substantial under-ascertainment of symptomatic cases, particularly at the peak of the first wave of the SARS-CoV-2 pandemic, in many countries.

Reported case counts will therefore likely underestimate the rate of outbreak growth initially
and underestimate the decline in the later stages of an epidemic.

Although there was considerable under-reporting in many locations, our estimates were consistent with emerging serological data,

suggesting that the proportion of each country’s population infected with SARS-CoV-2 worldwide is generally low.

And that's very good news about local authorities finally getting the data they need @BigChoc Badly overdue, and I do hope they pass the details on in real time as they get them, not weekly. I hope they also give them details of the locations where the people who test positive have been, as that's the kind of thing that local knowledge would really help know the significance of.

The UK and other countries in Europe have seen excess deaths fall to normal before lifting lockdown,
i.e. it's not lockdown causing most of the excess deaths, as is occasionally claimed

The ONS has found that a chunk of excess deaths not officially COVID were either COVID, or COVID appears a factor in worsening existing conditions that caused death

CalmYoBadSelf i think the two things can be reconciled. Based on the information provided by the ONS, over March, April and May the numbers with Covid, or suspected Covid, were higher than the reported figures, and for a short while excess deaths were significantly higher than that. This was when too few people were being tested, especially in care homes.

The position in the latest ONS weekly report is very different, with no excess deaths and the number with Covid as the cause or suspected cause of death actually less than what was reported over that period. So it is possible there is an over reporting problem, due partly just to very late reporting of historic deaths, and possibly due to the new issue identified of people who tested positive many months ago and have now died.

Yes, Firefliess track & trace by local public health & environmental officers, who know their patch is very effective

• it is what Germany has been using very successfully since the start of the epidemic

The UK has cut the number of these officers over the years, but there should still be enough to make a considerable local contribution

@CrunchyCarrot

Do we have any reliable data on the percentage of false negative Covid-19 tests? (Antigen tests, not antibody)

Home testing is likely to be particularly inaccurate, compared to tests done by professionals.

This NHS document suggests around 30% false negative tests

https://www.ouh.nhs.uk/working-for-us/staff/documents/staff-testing-privacy-statement.pdf

The test is a swab of your nose and throat undertaken at one of the Trust testing sites.
The test is not 100% accurate;
at present it is thought to detect approximately 70% of coronavirusi infections.
A negative test does not mean that you stop self-isolating if you have coronavirus symptoms.

FDA: Coronavirus Testing Basics

A useful overview of all testing methods, including swabs:

https://www.fda.gov/consumers/consumer-updates/coronavirus-testing-basics

•	You may have the virus, but the swab might not collect it from your nose or throat.
•	The swab or mucus sample may be accidentally contaminated by the virus during collection or analysis.
•	The nasal or throat swab may not be kept at the correct temperature before it can be analyzed.
•	The chemicals used to extract the virus genetic material and make copies of the virus DNA may not work correctly.

PHE report on the most recent deaths - those figures include people who die after a positive Covid-19 test. Since some people tested positive weeks ago there is a debate about how many died of Covid-19 and how many recovered and died of something else. These figures are the ones most often reported in the papers. Other countries stop counting people as Covid deaths if their death takes place more than 28 days after a test.

ONS publish, a little later, figures on total deaths in the uk. The difference between these figures and a "normal" death rate is called excess deaths. While there is a lot of debate about that too these are deaths rational people attribute to the effect of Covid-19, either directly or indirectly. The debate is about how many of these deaths were undiagnosed Covid-19 and how many are deaths because people did not seek, or were refused, medical care. It is far less clear cut than it seems since Covid-19 causes blood clotting and may have been the cause of deaths that appeared to be heart attacks and strokes. It may also be the cause of deaths recorded as pneumonia or even Alzheimers, since there is a genetic link suggesting Alzheimers sufferes may be more susceptible to the virus. During lockdown less travel meant fewer traffic accidents, cleaner air and less opportunity for young adult stupidity (like tombstoning) causing fatalities, it was not all negative. It's also debatable what time period you use when deciding what was an "average" year, since some include flu epidemics.

The statistics show that the NHS continued to provide urgent medical care even at the peak of the epidemic and even provided some non-urgent care. However when you underfund health care services and capacity has to be rapidly rejigged to care for virus patients you dont have intensive care beds and theatres available to do elective surgery. You dont have labs that can test samples from other patients if they are full of tests for the virus. You cant x-ray if all your capacity is taken up checking the lungs of patients because tests are not reliable. Your staff numbers are down considerably because PPE shortages mean a significant proportion are sick while the lack of testing for medical staff meant others were isolating when they or family members had Covid like symptoms but not Covid. And sick people held off going to hospital for fear of contracting the virus there. There will have been some deaths because people didnt get care.

It is worth pointing out that some commentators have very clear political agendas. The truth will be established in due course by proper research. Those with political agendas will lie and hope that lies repeated often enough become accepted.

Thank you to the poster who referred me to rp131 to understand the daily rates. I’ve also been looking for the historical prevalence estimates for infection rates - if Zoe tells me there are 2103 cases a day, I’d really like to know approx when we got to this on the upward curve.

I have really struggled to understand the govt policy throughout and it makes me very uncertain about the risk given the level of general social interaction there is over the past couple of weeks. It’s one thing if there were 2103 daily infections in say the second week of January; quite another if it was the last week of February, say.

Can anyone recommend a data source?

Thanks BCF. My DP had a Covid-19 test at a drive through center yesterday (but he had to self-test himself), tested on Day 2 of symptoms. He had a temp of 102'F yesterday but no cough nor loss of taste/smell. We thought it best for him to get a test. He felt pretty rough for a couple of days and sweated it out at night, today his temp is falling and he's feeling a lot better. Got a negative result. I am glad but can't help feeling a niggle of doubt. However I did read elsewhere that it also pays to consider the local risk. He hasn't been in contact with anyone who has Covid (we've only really seen one other person occasionally for the past several weeks) and this isn't a local 'hot spot' so probably it really is a true negative. But even so... I wish testing was more accurate. Perhaps the saliva test being trialled in.. Southampton? I think, may be rolled out if a success.

I found this article around the reliability of testing interesting

www.theguardian.com/world/2020/jul/05/risks-r-numbers-and-raw-data-how-to-interpret-coronavirus-statistics

It is worth pointing out that some commentators have very clear political agendas. The truth will be established in due course by proper research. Those with political agendas will lie and hope that lies repeated often enough become accepted.

It's true that commentators have very clear political agendas.

For example this article attempts to conflate the US and UK

time.com/5861697/us-uk-failed-coronavirus-response/

It correctly observes that the UK's excess death total is disastrous.

It then proceeds to try and blame this on 'track and trace' including an article on app developed in June, long after covid-19 had been brought under control in the UK by means of a lockdown.

It is not necessary for us to wonder whether the truth will be established in due course because we already know NOW that infections grew exponentially in March and then fell after lockdown very sharply. www.mrc-bsu.cam.ac.uk/now-casting/

It is a fundamental property of mathematics as true as 1 + 1 = 2 that if you have, say, 3 million people infected, then you cannot simply reduce that to zero over night, but rather there will be a reverse geometric progression where that declines from 3 million to 2 million to 1.5 million etc.

So points about poorly coded apps in June are irrelevant and the hallmark of someone more concerned with trying to score points than actually answer the question of 'why the UK has such a bad deaths total'.

Because we can see very clearly from statistics such as:

• infection estimates
• death counts (which follow infections)
• antibody surveys

that R numbers were very low after lockdown.

It follows that where a (the number currently infected) is 2 million, then a very hard lockdown, where r is say 0.3, is much more useful than when a is around 20,000 as now, since a fairly static % of a will die, so when a is a large number, r must be made much smaller to reduce that as quickly as possible, and where a is already small, then it would be ok for to be much closer to 1 than zero, because we do not have any policy of 'zero deaths in life', be it for covid-19 or for anything else.

The article then links to an article with a list of racist demands by advocacy groups

• It should be mandatory for NHS trusts to treat ethnic minority staff as “high risk and vulnerable” in regard to covid-19
• Ethnic minority staff should be redeployed away from covid-19 areas in hospitals wherever possible

and demanded that the report whose scope had already been made clear a month earlier, focus solely on ethnicity and ignore more important factors such as sex and obesity.

www.gov.uk/government/news/review-into-factors-impacting-health-outcomes-from-covid-19

Dated 4 May 2020

The exercise is part of a rapid review being led by Public Health England (PHE) to better understand how different factors such as ethnicity, deprivation, age, gender and obesity could impact on how people are affected by COVID-19.

Anyway.....

The covid-19 excess death total in the UK was arrested once the UK shutdown. This happened after an explosion of infections and ever since then infections and deaths have been going down continuously.

This is quite different from the US response which has been a slow-motion disaster for months on end.

The UK might have ignored science in January - March. A review will find that out. The consequences of not acting earlier were disastrous.

However we know why the UK has such a high death total, and that is because it failed to arrest the viral growth in time.

As for 'establishing the truth following proper research', that is possible, but it is not clear that any such 'truth' will be accepted.

We can now pick whichever truth suits our pre-existing prejudices. That will not change. The media has its prevailing themes and amplifies the voices which support them. Things which don't fit in with the news narrative will be discarded.

I am hopeful and have reason to believe that the UK has learned from its disastrous outcomes.

Other countries I have much less hope. The US response has been very poor, and a number of developing countries will continue to be hampered by poor levels of education, belief in quack cures, and then as they experience the consequences of these, the lack of a tax and GDP base with which to address the required healthcare and testing response.

@BigChocFrenzy Not sure if you read the thread as it is talks about it being comparable to 1957 flu or 1918 pandemic not the more recent endemic flu.

Other countries stop counting people as Covid deaths if their death takes place more than 28 days after a test

So how would this work for someone, say, like Kate Garraway’s husband? He has been hospitalised for months now, is COVID free but is still desperately ill. If he were to die now, how would his death be recorded? Both in the U.K. and in countries imposing a 28 day limit?

@Howaboutanewname he would be a Covid death as he would have Covid on his death certificate, and he is in hospital.

The deaths that are being questioned are those that are recorded as happening outside of hospital, more than 28 days after a test, and without Covid on the death certificate.

underlying trends in longevity

This came up last night in a chat with friends. UK & USA life expectancy has already stalled, and indeed was falling, prior to Covid. Whether this has any significance with regards to Covid deaths remains a mute point.

Why the government didn't move to the "local council preparedness for civil crises" much, much earlier, is another big question. Councils have always had in place these levels (Gold being the highest) which enables them to put in place (with Police, Army and Health Services) local responses to crises/disasters. This much more closely mirrors the German district response model, which of course has been so successful. Some councils did implement, as far as they could, these plans, and were very successful in delivering local help during the lockdown for example, but, as we have seen with the recent announcement, with access to personal medical data, couldn't do anything further. This seems to be our downfall. We do have systems in place, in both NHS and local councils, but it very much depends on the management and resiliance of both (both in terms of management and finance) in order for them to effectively respond.

without access to medical data

"It should be mandatory for NHS trusts to treat ethnic minority staff as “high risk and vulnerable” in regard to covid-19"

I'd have expected the oldest staff to be protected
The BAME increased risk, once we normalise for occupation etc doesn't compare with the risk from age,
which increases about 10x for every 20 years of age

So e.g. an HCP aged 62 has about 100x the risk (not 100 %, but 10 x) of one aged 22

Similarly for order of vaccination, when one comes
Of course frontline HCPs and emergency services come first, but then I'd expect the age 80+ group, then 70+ etc before all but the most serious conditions
e.g. other conditions like T1 are "only" 3 x risk, which is like being maybe 7-8 years older, all other fators being equal.

[quote whatsnext2]@BigChocFrenzy Not sure if you read the thread as it is talks about it being comparable to 1957 flu or 1918 pandemic not the more recent endemic flu.[/quote]
...
Most of those differences apply just as much to the 1918 / 1957 pandemic type flu

I was making the point that pandemic plans in the UK & most other countries were for flu

• either a v severe strain, or a year where the vaccination wasn't as effective as normal -

which has important differences in characteristics to COVID

Hence why the plans worked so poorly