COVID-19 Data & Docs

[Grizzly]

https://twitter.com/robertmaguire_/status/1249072051207503874?s=21



Wish I could embed this video of Obomber, but here's JR the lesser



Watch both at the original twit link..

[alloneword]

Thanks for that Jack (& Griz!) - a lot to think about...

excess deaths, i.e. total deaths from all causes compared against expected numbers based on the past, are surely the most reliable indicator of C19 impact whether direct or collateral.

Totally agree.

The problem is distinguishing between those 'direct' and 'collateral' deaths, so my focus now is on identifying reliable data regarding the effectiveness and consequences of the 'lockdown'.

There's a chap called 'Yinon Weiss' who's had a fair stab at the US data, here:

How effective are State wide shutdowns versus more measured approaches? That is what I will analyze in this post by looking at the data.

(He plots variations of 'time from 'lockdown'' against 'x per million deaths').

https://medium.com/@yinonweiss/lets-vis ... 3a5cdb50ad

There are a few others looking at similar angles, I'll try to sift the better ones out and post here.

[alloneword]

A very good piece on Sweden & stats:

..it is important that some of the simplistic thinking surrounding this crisis needs to be challenged. If the only moral and virtuous policy is to minimise Covid deaths at all costs, as parts of the media imply with the endless game of ‘look how badly Britain is doing compared to X other country’, there can be no argument for releasing any part of the national shutdown at any point. The refrain at every point will be the same: why take a risk with people’s lives? And the Government will find itself backed into a corner.

A better definition of success would surely be: which governments are getting the balance right between protecting their people as reasonably as possible against this new threat while not destroying too much of their country in the process? Death rates per million is not the only datapoint in this difficult equation; and right now, it is far too early to judge how successful Sweden, or the UK, will be.

https://unherd.com/2020/04/jury-still-o ... -strategy/

[alloneword]

Latest UK 'all cause' stats: https://www.ons.gov.uk/file?uri=%2fpeop ... 42020.xlsx

Showing a ~60% increase on 5 yr average overall, but with some oddities...

16,387 total (all cause), 5 yr av of 10,305 = 6,082 'excess'.

Of those, 3,475 'where COVID-19 was mentioned on the death certificate'... leaving 2,607 (43% of the excess) where it wasn't?

If that is somehow due to under reporting of CoV-19 on Death Certs, I'm struggling to see how that could be - if anything, the guidance given would tend to lead to the opposite.

Incidentally, the breakdown of 'place of occurrence' shows:

3.45% Home
89.50% Hospital (acute or community, not psychiatric)
0.89% Hospice
5.61% Care Home
0.09% Other communal establishment
0.46% Elsewhere

(This is only the 'CoV-19 deaths', so it tells us nothing about that 43%)

Also of note:

We will publish extra articles periodically, giving enhanced information such as age-standardised and age-specific mortality rates for recent time periods, and breakdowns of deaths involving COVID-19 by associated pre-existing health conditions.

I'll keep an eye out for that.

[alloneword]

New(ish) study from John P. A. Ioannidis, attempting to communicate 'risk' in understandable terms:

The COVID-19 death risk in people <65 years old during the period of fatalities from the epidemic was equivalent to the death risk from driving between 9 miles per day (Germany) and 415 miles per day (New York City).

https://www.medrxiv.org/content/10.1101 ... 20054361v1

(OMG! Driving 10 miles in Germany gives you Covid!?!?!1!?!)

[JackRiddler]

I can see the panic media producing that headline.

I don't have time to read it right now but does he deal with the actual argument for "flattening the curve"? Meaning, can valid data show that the very large population who do have the much higher risk (older, already suffering from other conditions) are equally protected by other means, without keeping everyone else on lockdown? Because notwithstanding the extra panic-mongering, that's the rationale of the medical authorities guiding the policy.

[alloneword]

He stresses that:

..no matter what strategy is selected for addressing COVID-19 in the current or future epidemic waves should include special emphasis in protecting very elderly individuals.

[p.16]

and that 'lockdowns' constitute a

fully justified “better safe than sorry” approach in the absence of good data.

However, there is a 'however' to that.

He references Wittkowski (his [30], my link works) as arguing that:

lockdowns may be even harmful as a response to COVID-19 itself, if they broaden rather than flatten the epidemic curve

But yes, the argument for lockdown is made 'in the absence of good data'. And if policy is not being guided by 'data' or 'science'..?

What I find strange now is that Ferguson in fact considered "Social distancing of those over 70 years of age" in *that* 16th March NPI Modelling paper (he labelled it 'SDO'), described as:

Reduce contacts by 50% in workplaces, increase household contacts by 25% and reduce other contacts by 75%. Assume 75% compliance with policy.

[T.2, p.6]

But then appears to have treated it as a sort of 'last resort' measure, so that it never appears in isolation within the results of the modelling (or at least, those published), only in conjunction with or in addition to all other mitigation measures, before seemingly disappearing from consideration altogether.

[alloneword]

The UK ONS 'extra articles' mentioned above should appear on Thursday, here: https://www.ons.gov.uk/releases/analysi ... ingcovid19

[alloneword]

Factors associated with hospitalization and critical illness among 4,103 patients with COVID-19 disease in New York City [preprint]:

https://www.medrxiv.org/content/10.1101 ... 20057794v1

[alloneword]

Exposure to air pollution and COVID-19 mortality in the United States [preprint]:

https://www.medrxiv.org/content/10.1101 ... 20054502v1

[Grizzly]

Gain-of-Function Research: Ethical Analysis
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996883/

[alloneword]

Some ONS datasets on the social aspects/impacts in the UK:

Coronavirus and the social impacts on Great Britain data

And 'business':

Business Impact of COVID-19 Survey (BICS)

[alloneword]

Latest UK 'all cause' stats: https://www.ons.gov.uk/file?uri=%2fpeop ... 42020.xlsx

Showing a ~60% increase on 5 yr average overall, but with some oddities...

16,387 total (all cause), 5 yr av of 10,305 = 6,082 'excess'.

Of those, 3,475 'where COVID-19 was mentioned on the death certificate'... leaving 2,607 (43% of the excess) where it wasn't?

[...]

We will publish extra articles periodically, giving enhanced information such as age-standardised and age-specific mortality rates for recent time periods, and breakdowns of deaths involving COVID-19 by associated pre-existing health conditions.

I'll keep an eye out for that.

The UK ONS 'extra articles' mentioned above should appear on Thursday, here: https://www.ons.gov.uk/releases/analysi ... ingcovid19

Which it did. Very detailed in terms of a full listing by ICD-10 code (cause of death), but only for the month of March as a whole - a time period of which they state:

The overall mortality rate in March 2020 was lower than the five-year average

So this data will perhaps serve as a baseline (I have found no other that classifies by ICD-10 coding) against which to compare future data (April? - ONS repeatedly reference 'each month' in the description), hopefully gaining some insight into what could have caused that 43% of the 'excess' deaths in week 14 that were not attributed to Covid-19. Here's hoping that (we and) the format can survive another 4 weeks.

The practice of publishing some dataset grouped by week, while others are grouped by month presents challenges, making comparison difficult. Also, declining to publish separate figures for ICD-10 codes U07.1 and U07.2 is interesting:

An emergency ICD-10 code of ‘U07.1 COVID-19, virus identified’ is assigned to a disease diagnosis of COVID-19 confirmed by laboratory testing.

An emergency ICD-10 code of ‘U07.2 COVID-19, virus not identified’ is assigned to a clinical or epidemiological diagnosis of COVID-19 where laboratory confirmation is inconclusive or not available.

Both U07.1 and U07.2 may be used for mortality coding as cause of death

https://www.who.int/classifications/icd/covid19/en/

There are 71 ICD codes on that ONS list, so I'm not sure one more would have made a difference.

[alloneword]

Oxford University - The Centre for Evidence-Based Medicine: Oxford COVID-19 Evidence Service

"Rapid reviews of primary care questions relating to the coronavirus pandemic, updated regularly" (76 articles)

https://www.cebm.net/oxford-covid-19-evidence-service/

A particularly useful article: Reconciling COVID-19 death data in the UK

Some further thoughts on the difficulties presented when interpreting ONS data: https://hectordrummond.com/2020/04/08/t ... -category/

The ONS have a blog - https://blog.ons.gov.uk/ - and a Twitter account - https://twitter.com/ONS

Also: Social isolation increases death risk in older people (study: https://www.pnas.org/content/early/2013 ... 51adbcbabf )

[alloneword]

Another interview with Dr John A Lee:

It seems incredible to me that we are not equally as interested in the effects of the lockdown on lives and livelihoods as we are in the actual virus itself. I think we are guilty at the moment of being a bit monomaniacal and focusing only on one thing, and really not focusing enough on the consequences that are coming out of what we have done to face this one thing.

https://www.spiked-online.com/2020/04/1 ... e-working/