At 20:19 06/03/2020, Thomas Cox wrote:
>So, by the time the WHO actually issues a pdf, I suspect that the
>data incorporated in it is quite old in pandemic time as I am sure
>the draft document is reviewed by many parties before it is released.
Indeed, and that is fully acknowledged. The WHO data has a cut-off
time of 09:00 UTC, but is not published until much later (around
22:00 UTC) on the day.
It has really become an essentially unimportant and almost 'semantic'
issue, perhaps partially of my doing. I suppose because it is
published (much later) on the same day, I have been describing the
the data with a cut-off of, say, 09:00 UTC on 05 March as being "05
March data". If I had, perhaps more reasonably, described it as "04
March data", then it would be almost identical to what JFU are
calling 04 March data, and discussions about the apparent
discrepancies would never have arisen. I suppose it is, in some
senses, a 'rounding' issue. However, as I think we are agreed ...
>However, as you note, in an exponential growth model a days lag is
>almost irrelevant, the general shape of the curve, and the levels
>one would be likely to forecast for the next couple of time cycles
>are unlikely to differ greatly.
Indeed so. In fact, as above, if one 'rounds' one description of
what day the data relates to, then the two sets of data are identical
for any day - the only real difference being that the JFU data which
actually relates to a particular day is published much earlier. If
the growth of cases is following a modelable curve, it really makes
little difference which one is used for forecasting.
However, forecasting is not something I am inclined to get into,
since we need proper epidemiologists here. If we approached it
simply as a mathematical exercise, we would be assuming that an
exponential rise would continue indefinitely all the way to the point
at which the entire world's population has been infected. However,
that's not how epidemics, even pandemics, behave in practice - so
I'll be interested to learn a bit about how they do set about
modelling epidemics.
>So, for example, the JHU data might incorporate a report from a
>trusted person with contacts in a hospital or a nursing home, who
>might get a report of a new case, that has not been released in a
>public statement or by an official public health information provider.
That's the sort of process about which I was previously expressing
concern. I find it hard to see how they could establish with any
great certainty that 'a report from a trusted person' really did
relate to a case that had not already appeared in the 'official'
figures. Whilst I understand that the usual privacy/confidentiality
considerations have been relaxed, I very much doubt that the official
figures (now relating to hundreds or thousands of cases each day)
are flying around the world with attached lists of the names,
addresses, DOBs etc. of all the cases!
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]
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]
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.
*******************************************************
|