Also Spain (using lock down date = 2020-03-13) , France (using lock down date = 2020-03-18) and the US state New York (using lock down date 2020-03-23) are in a very similar situation (scale factor a bit above 1) as Italy. They are just lagging 6-16 days behind Italy and in the end will have the same number of deaths or more.
Comparison New-York (lock down date= 2020-03-23) with Italy (lock down date = 2020-03-07).
I didn't scale it (as scale factor = 1.027).
Observations
- death charts are very similar (only shifted 16 days).
- New York has more than double the number of confirmed cases compared to Italy 16 days ago (this might be due to more broadly testing).
Comparison New Jersey ( lock down date = 2020-03-25) with Italy ( lock down date = 2020-03-07).
Observations
- About 18 days behind the path of Italy (based on best fitting chart
total deaths (limited to 7 days in future)for several different lock down dates). - Total deaths is about 30% of total deaths in Italy 18 days ago (note that the Y- values of the charts are scaled for New Jersey).
Comparison US (all states) ( lock down date = 2020-03-23) with Italy ( lock down date = 2020-03-07).
Observations
-
date shiftof 16 days with Italy gives best fit of charttotal deaths (limited to 7 days in the future) - Situation in the US is about 2.35 times worse (in terms of deaths) compared to Italy (16 days ago).
- The number of deaths and confirmed cases is decreasing in the US over the last 3 days. The numbers in Italy didn't decrease that early and that fast. Maybe this decrease is caused due to the weekend (and Eastern) resulting in less reported cases ?
- The US reported 2.5 times more confirmed cases per death than Italy. This might be explained due to more broadly testing in the US compared to Italy (16 days ago).
Worrying indeed.
Do you have new analysis of UK from latest figures? Is the trajectory same/better/worse?
Comparison of UK (lock down date = 2020-03-23) with Italy (lock down date= 2020-03-07).
Observations
- The main trend is still the same as 10 days ago (current
scale factor= 1.130 while it was 1.211 ten days ago). - The last 3 days in a row the number of deaths decreased in the UK (something Italy didn't experience around the same period). So UK might be over its peak but it will still take weeks before numbers will really go down if we compare with Italy.
1 Like
comparison of Italy (lock down date = 2020-03-07) with Hubei (lock down date = 2020-01-23)
Observations
- So Italy is already 18 days beyond the peak of death (= around 28 of March) but it is clearly not descending that rapidly as in Hubei.
- The number of confirmed cases is descending very slowly. The chart is very unlike the chart of Hubei.
- If Italy would have followed the same trend as Hubei then within 18 days (=3rd of May) the number of confirmed cases would go down to 0 !
Main conclusion:
Italy and most likely most other countries will not succeed in controlling the virus as good as it was done in the Province Hubei where the outbreak began.
Don't know if you saw the news, China just updated the deaths in Wuhan by 50%
Thanks for sharing this. I also think that the actual deaths in Hubei might be considerably higher than reported as they reported considerably more confirmed cases per death if we compare this with other European countries that are similarly affected. ... but I don't think this news invalidates the comparisons made with Hubei as I am scaling them. (of course the real scale factor (or what is called hubei factor in earlier charts) should be lower).
What might invalidate these comparisons if the number of deaths and confirmed cases didn't really go down like they have reported but this claim is not made in the above news link. Note that for Hubei only 3 new confirmed cases were reported since last 30 days (started counting from 2020-03-19) and during the same period only 100 people died (according to their reports) from covid-19.
It must also be noted that other countries are as well not counting deaths that are dying from covid-19 or presumably covid-19 outside the hospital (in fact I think that Belgium sofar is the only country that is counting those people in their stats and partially due to this has the worst stats of all countries).
Scotland is including Care Homes at least. But this is indeed a major issue. Though to be honest, we won't really get a clear picture until we can compare year on year.
1 Like
Netherlands appears to only report deaths for people dying in hospitals who got tested positive. Care/nursery homes arenât included as most people there simply arenât tested just assumed they have it and when they die itâs not logged in the publisher statistics. Same for people dying at home. The statistics institute is suggesting the deaths are twice as high on a weekly base than what is reported.
For these charts that simply means the Dutch stats really canât be used for any comparison
1 Like
I have released a new version of my flow on github with following main feature:
- when
smooth chartsis set for locationChina:Hubeiit will also ignore the 1290 extra deaths and 325 extra confirmed cases reported at 17 April 2020. I did this as these deaths and confirmed cases didn't really happen at that date.
Here below you see for location China:Hubei both the charts with smooth set (= blue) and not set (= red).
In yesterdays news Trump claimed to expect about 65,000 deaths in the US.
He must have looked at my charts.
See here below US (lock down date= 2020-03-23) versus Hubei (lock down date = 2020-01-23).
So if US follows same trend as Hubei then they will end up with 65,000 deaths.
... but it is unlikely that US or European countries will follow the same trend as Hubei. The number of deaths / confirmed cases once over the peak are not that rapidly decreasing as reported in Hubei (this you can already clearly see when comparing Italy with Hubei) so 65,000 deaths for the US is a too optimistic view.
It would be interesting to see the China data with the new figures applied as a retrospective average.
At some point the deaths for England and Wales will likely double as care home and other deaths are included (if they ever are).
I think that the most interesting country for most of us to compare against now will be Germany. They had early large-scale testing and plenty of ICU beds and are most likely giving the most accurate data out as well.
3 Likes
If you mean that the correction of 1290 extra deaths reported at 17 April should be retroactively applied to the past proportional to the number of deaths reported at those days then this just means that we need to add 40% of the deaths (or scale it with a factor 1.4).
Here below the comparison between Germany (lock down date = 2020-03-17) and Hubei (lock down date= 2020-01-23).
Observations and comments
- I have smoothed (= weighted 3 days averaging) the German data as they reported very fluctuating deaths (even negative new death totals) and new confirmed cases.
- Relatively (scaled based on total number of deaths):
2.1 Germany is reporting considerably more confirmed cases during the beginning (around the lock down date) than Hubei which is not unexpected due to the more extensive testing in Germany around that period.
2.2. During the peak of confirmed cases (around 1st of April) the relative difference between Germany and Hubei is rather small so this might indicate that testing in Hubei is comparable to Germany.
2.3. After the peak of confirmed cases (around 16th of April) the relative difference between Germany and Hubei is again high and this might indicate that the number of cases in Germany is not dropping as fast as in Hubei.
Or that the Chinese numbers are a fabrication?
charts for the 2 most affected Chinese provinces (in terms of number of death) outside Hubei
- Heilongiang (population 38 million)
- Henan (population 96 million)
Observations
- Although the number are only a fraction of the numbers of Hubei the charts are looking very similar (peaking at the same time, going down at the same time).
- For Heilongjiang we see a second peak of confirmed cases around 12th of April which might be caused or triggered by the import of new cases from Europe which were peaking around that time.
... after looking a lot at the charts, I wouldn't classify the Chinese numbers as a fabrication.
There might be an underreporting of the number of deaths (at 17 April they have made themselves a correction of 40% for province Hubei). I especially think that for the other provinces they have underreported the number of deaths as the number of deaths compared to the number of cases is too low.... but this underreporting doesn't really matter for the comparisons made in the thread. The charts show a relative (scaled) comparison - so if the actual death numbers in Hubei should be multiplied by 3 then this doesn't change the scaled charts or the extrapolated number of deaths.
Of course you could also claim that the relative drop in confirmed cases and deaths is fabricated
but that would mean that China is/was still having many confirmed cases and many deaths while they were reporting zero deaths and confirmed cases (or close to zero) for several weeks. I don't believe you can hide high numbers of confirmed cases and deaths so long for the world and I have also seen no trustable news source claiming so.
3 Likes
And here we go ... UK Just announced that its death rates are 40% too low.
I'm also getting other feedback that hospitals are deliberately avoiding testing people unless they go into ICU.
1 Like
It appears to be rather normal that most COVID deaths outside hospitals are not reported in the COVID country statistics. Belgian is an exception on this (that is also why we almost ranked number 1 in terms of Deaths / 1 Million population).
FYI in Belgium:
- 47% died in Hospital
- 52% in eldery care (of which 4% are confirmed and 96% are presumed covid deaths).
- 0.4% at home
- 0.4% on another place.
So as you can see the figures for Belgium doubled when counting also the presumed covid deaths in eldery care.
1 Like
