hi Diana, all
the JHU time series data is on github here:
https://github.com/CSSEGISandData/COVID-19/tree/master/csse_covid_19_data/csse_covid_19_time_series
But I think your suggestion won't work in the UK, because the "cases" in
this data are just those cases confirmed (mainly in hospital) by testing.
The whole point is to estimate the cases in the full population. Maybe it
would work in China though, if surveillance was complete there.
let's look at the Hubei data, attached
on the "data" tab I've put the figures from github, for cases and deaths
in Hubei
on the "regress" tab I've regressed deaths on cases, with a lag, and
reported R2 and the coef showing dependence of deaths on lagged cases.
I've done this in Libre Office (i.e. Excel equivalent) by hand so there's
no macros, coding or R. Hopefully I've done it correctly!
What happens is that R2 peaks at a 6 day lag, and the death rate is just
under 4%.
However, this does not mean Pueyo is wrong in saying death is 17 days
after infection. It could just mean Hubei cases were identified 11 days
after infection.
Anyway, even if we go back to the UK and assume the real death rate is 4%,
not 1%, the estimated cases on 1 Mar drop from 10,000 to 2500 and the
estimated cases on 15 April are likewise 4x lower at 1.8 million, but as
the death rate is 4x higher, we would still expect 72,000 cumulative
deaths by 2 May.
Greg
> Hi
>
> Pueyo's estimate
> PLEASE. Where can one download the JHU .csv files?
> Presumably they have daily new cases and daily deaths by time
> Then one can estimate new deaths (t)/(new cases(t-d) for different
values
> of delay d, separately for each country.
> Don’t think the general message of UK current cases between 250k and
1000k, best estimate 750k
> Data sources, Ones I find most useful
> JHU
> https://www.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6
UCL http://nrg.cs.ucl.ac.uk/mjh/covid19/
> Both updated daily, but may be behind some news announcements
> Imperial college’s document - the’ Science' on which UK government
actions
> are supposedly based can be downloaded from
> Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf<https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf>
Static, so not updated. BUT has
> 1. Estimates based on replication rate R0 and trigger events for
interventions
> Assumes that interventions are fully implemented. Different if
‘mandatory
> (France) or advised (UK), cf Johnson were extolling his local
> 2. For cases identified: hos[pitalisation, icu and death rates.
> Found this useful, Know how likely I am to die if I get infected! A
depressing 1/20
> GAME CHANGER Antibody test
> If there is antibody identifier from blood within 1 month??
> We could have whole communities of immune folk.
> E.g. schools for immune
> We could have only immune NHS and social career people working
> We could have testing stations next to. Security at airports
> We could junk all the current estimates and start anew
> BUT weight NEVER attain herd immunity of 60% (not that this was ever
realistic, see Imperial College estimates
> ADVICE?
> Why is it risky to walk/run in deserted streets or parks?
>
> best
> Diana
>
>
>
> On 19 Mar 2020, at 00:55, John Whittington
> <[log in to unmask]<mailto:[log in to unmask]>> wrote:
>
> At 21:06 18/03/2020, Greg Dropkin wrote:
> If so, there are now around 10,000 * 2626/35 = 750,000 actual UK cases.
Do
> we think so? If not, which bits of this are wrong?
>
> I can't fault your arithmetic (on the basis of the assumptions you've
used) and nor do I think that 750,000 is beyond credibility for the
cumulative number of people infected in the UK.
>
> However, per my various other mutterings, I am increasingly uncertain of
what the 'UK numbers' we're seeing actually mean and represent.
>
> One hopes/imagines that most/all people admitted to hospital with
suspected Covid-19 infection will be tested, so the figures we're seeing
ought to at least include most test-confirmed cases who have been
hospitalised, but goodness knows what else they include. For example, at
> the extreme, if much of the testing were of the population in general
then, a 1.3-fold day-to-day increase in reported 'cases' (positive tests)
> might merely mean that the number of tests undertaken was increasing by
that factor every day!
>
> Is there any information available on the number of hospitalisations for
proven Covid-19 infection? If it were, that might give a much more
reliable indicator of what is going on than total figures for 'positive
tests' (in a population of unknown nature).
>
> Kind Regards,
>
>
> John
>
> ---------------------------------------------------------------- Dr John
Whittington, Voice: +44 (0) 1296 730225
> Mediscience Services Fax: +44 (0) 1296 738893
> Twyford Manor, Twyford, E-mail:
> [log in to unmask]<mailto:[log in to unmask]>
> Buckingham MK18 4EL, UK
> ----------------------------------------------------------------
>
> ******************************************************
> Please note that if you press the 'Reply' button your
> message will go only to the sender of this message.
> If you want to reply to the whole list, use your mailer's
> 'Reply-to-All' button to send your message automatically
> to [log in to unmask]<mailto:[log in to unmask]>.
> Disclaimer: The messages sent to this list are the views of the sender
and
> cannot be assumed to be representative of the range of views held by
subscribers to the Radical Statistics Group. To find out more about
Radical Statistics and its aims and activities and read current and past
issues of our newsletter you are invited to visit our web site
> www.radstats.org.uk<http://www.radstats.org.uk>.
> *******************************************************
>
> ____________
> University of Hertfordshire
> College Lane, Hatfield, Hertfordshire AL10 9AB, UK
> +44 (0) 208 444 2081
> +44 (0) 7403 18 16 12
> [log in to unmask]<mailto:[log in to unmask]>
> http://dianakornbrot.wordpress.com/
> http://go.herts.ac.uk/Diana_Kornbrot/
> skype: kornbrotme
> Save our in-boxes! http://emailcharter.org
> __________________
>
>
>
>
>
>
>
> ******************************************************
> Please note that if you press the 'Reply' button your
> message will go only to the sender of this message.
> If you want to reply to the whole list, use your mailer's
> 'Reply-to-All' button to send your message automatically
> to [log in to unmask]
> Disclaimer: The messages sent to this list are the views of the sender
and
> cannot be assumed to be representative of the range of views held by
subscribers to the Radical Statistics Group. To find out more about
Radical Statistics and its aims and activities and read current and past
issues of our newsletter you are invited to visit our web site
> www.radstats.org.uk.
> *******************************************************
>
******************************************************
Please note that if you press the 'Reply' button your
message will go only to the sender of this message.
If you want to reply to the whole list, use your mailer's
'Reply-to-All' button to send your message automatically
to [log in to unmask]
Disclaimer: The messages sent to this list are the views of the sender and cannot be assumed to be representative of the range of views held by subscribers to the Radical Statistics Group. To find out more about Radical Statistics and its aims and activities and read current and past issues of our newsletter you are invited to visit our web site www.radstats.org.uk.
*******************************************************
|
