Coronavirus Analysis (part 1) - How long will lockdown last?

in #coronavirus4 years ago (edited)

Some very basic analysis based on the figures available from wikipedia.

CHINA
NUMBER OF NEW CASES AND LOCKDOWN

Here's a chart of new cases in China by date. The outbreak was centred on Wuhan / Hubei with 80%+ of cases coming from this area.

China CV3 2903.png

China new coronavirus cases by date
Figures sourced from:
https://en.wikipedia.org/wiki/2019–20_coronavirus_pandemic_in_mainland_China

Lockdown started in Wuhan on 23 January and other Hubei cities on 24 January. There was an exodus of people from Wuhan just before the lockdown which may have weakened the quarantine and allowed further spread of the disease.

Looking at the new case numbers there is an initial levelling around the 11-14 day mark (after lockdown) followed by a general decline in cases with some later spikes. These spikes may be down to issues with changes in the approach to measure the number of cases or possibly the imperfect quarantine noted above.

New case numbers fell below 100 on 6 March. The lifting of the lockdown in Hubei started on 13 March, 49 days / 7 weeks after the lockdown started. The lifting of the quarantine in Wuhan is happening now, representing 9 weeks of lockdown.

NUMBER OF SICK PEOPLE

It's worth also looking at this chart, sourced directly from wikipedia:

China number of sick people.png
China number of sick people by date
Figures sourced from:
https://en.wikipedia.org/wiki/2019–20_coronavirus_pandemic_in_mainland_China

Interesting points to note are:

  • Although new case numbers had fallen to 200 by 1 March, the number of sick people at this point was 33,000 and there were still 10,000 sick people on 15 March, 2 weeks later. I think this has implications for the length of any lockdown.
  • The peak in numbers of sick people (and deaths) appears to lag the peak in the number of cases by about 1 week. So the heaviest strain on health services will continue for a week after the lockdown has caused a peak in new case numbers.
SUMMARY

So a timeline summary based on the above:

  • Lockdown period to peak of new cases: 2 weeks
  • Lockdown period to low number of new cases: 5 weeks
  • Lockdown period to potential lifting of quarantine: 7 weeks (although many people still sick in hospital at this point)
  • Full lockdown in hardest hit areas: 9 weeks

China's quarantine was probably pretty well enforced. So applying these figures to other countries assumes the same stringency of lockdown. On the other hand, China suffered the virus first and other countries should be able to learn some lessons and benefit from advances in technology, in particular testing, to assist in handling the outbreak.

In part 2 I'll look at some European countries and use the figures from China to project what happens next in these countries.

Sort:  

Nice one @miniature-tiger.

Like you said not sure will this numbers be the same for the west, as lockdowns are not as popular. From what we are already seeing it probably will take longer. Italy one example of that.

I've got some numbers for Italy and I'll post them up later. I think their lockdown is working. They are over the peak of new cases and the rate of new cases is similar to China at the same point in their lockdown.

This week is probably the worst for their hospitals given the lag between new cases and total sick patients but another couple of weeks of lockdown and they could be through the worst of it.

Very interesting analysis, got to go with the data, but as you say delays in testing will distort this! It takes several days ATM to get a result in the UK!

True. When looking at the UK numbers (and most likely other European countries - excluding Germany) I think it's worth noting that:
(i) The official "number of cases" is actually the number of cases with symptoms severe enough to be in hospital + Prince Charles + Boris + Premiership footballers. The actual number of cases, including asymptomatic cases, may be something like 20x as high, assuming 5% of cases develop severe symptoms (which is an estimate I heard early on).
(ii) The data is going to be ropey. Testing delays as you say, differences in testing approaches between countries, changes in measuring approach (UK changed over the last couple of days), differences in classification (tested, symptomatic etc). So all conclusions are very broad-brush.

That said, the numbers are sufficiently interesting for me to draw some high level conclusions - as I think the numbers are fairly encouraging in Europe and the media coverage is typically sensationalist.

The media coverage is a fucking joke. Breakfast Time and the One Show especially - never has there been a more compelling case for scrapping the damn license fee!

It's a fascinating sociological/ methodological issue for sure though.