explicitly, the unmitigated scenario, once again assuming the parameters
are right and I've set this up correctly, is that on day 136 (end of the
12 weeks) there would be 1.28m deaths (and not many after that). I can't
tell you the age distribution of deaths, as the model is pooled - some
day, maybe.
Greg
> Great stuff! I hope Boris sees this. But ...
> How does the 25k deaths compare with a 'do nothing' scenario?
>
> How many life years are lost?
>
> On Tue, 24 Mar 2020, 08:20 Greg Dropkin, <[log in to unmask]> wrote:
>
>> hi
>>
>> same model, but with a better graph. still assuming transmission from
>> mild
>> cases reduced to 10% of what it is currently, for 12 week lockdown.
>>
>> nb expect 25,000 deaths by end of lockdown
>> and expect severe cases (requiring hospitalisation) to peak at 29,000 on
>> day 62 = Apr 1, and critical cases (requiring ICU) to peak at 14,000 on
>> day 73 = Apr 12.
>>
>> Greg
>>
>> > hi
>> >
>> > so here's what happens if the lockdown reduces the transmission from
>> mild
>> > cases to 10% of what it is currently [a heroic assumption, I think],
>> and
>> the lockdown remains in place for the famous 12 weeks Boris mentioned
>> the
>> > other day, but not tonight.
>> >
>> > I've extended the vertical axis in the 3rd panel so you can see what
>> happens by day 80, four weeks from now.
>> >
>> > The lockdown does work, but around 13,000 people still die in the next
>> 4
>> weeks, and if the lockdown is simply turned off after 12 weeks and
>> transmission returns to the current levels the problem bounces straight
>> back. However, it does buy time and that is important.
>> >
>> > My bet is that they will not end the lockdown after 12 weeks, although
>> they might go for a partial easing for a bit.
>> >
>> > stay well
>> >
>> > Greg
>> >
>> >> hi
>> >> I expect people have seen stuff like this, but I've learned how to
>> run
>> > the
>> >> first stage of Alison Hill's SEIR model, which makes it possible to
>> set
>> > parameters to fit what we know re UK growth rates in the exponential
>> period.
>> >> attached is a chart and the R code for it.
>> >> we knew (last week) that the log growth rate is around 0.26. I've set
>> > the
>> >> transmission rate for infectious (mild) cases to try to achieve that.
>> I'd guess that whatever mitigation is now in place will not have
>> > affected
>> >> the deaths yet - as it is over 20 days from infection to death. look
>> at
>> the 3rd panel, which is a close-up of the deaths. We know there
>> > were 281 deaths as of 22 March. Therefore, from the graph, we are on
>> day
>> 52, i.e. the UK story began 52 days earlier, on 31 Jan - exactly when
>> they
>> >> first found it.
>> >> So I've added grey dotted vertical lines at day 52 in the other
>> panels.
>> > This shows you where we are, and what's just about to happen if the
>> mitigation doesn't work.
>> >> Unlike my previous comments predicting from exponential growth, these
>> > are
>> >> predictions from the SEIR model. But I've only modified the
>> transmission
>> > rate (mild) - leaving all other parameters at the defaults from alhill
>> site.
>> >> I'm not interested in interactive display (shiny etc) at this stage
>> > (maybe
>> >> never), just the model, model fitting, and parameter values in
>> relation
>> > to
>> >> published data incl clinical. And some day, the impact of mitigation.
>> please email me off list if you want to talk about the code.
>> >> stay well
>> >> Greg
>> >
>> >
>> >
>> >
>> >
>> > ******************************************************
>> > 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.
>> *******************************************************
>>
>
> ******************************************************
> 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.
*******************************************************
|