THE LAG PROBLEM IN OBSERVING CORONAVIRUS
There is a 3-9 days lag between contagion & hospitalization and another 5-10 days between hospitalization and death.
This poses severe challenges (thread)
There is a 3-9 days lag between contagion & hospitalization and another 5-10 days between hospitalization and death.
This poses severe challenges (thread)
2/ One challenge is that when computing mortality rates, we should compare today's deaths with the number of people with symptoms about one week ago.
Eg. today 304 deaths, 14,551 confirmed cases, ratio is 2.1%.
However, one week ago we had 1312 cases, mortality becomes 23%.
Eg. today 304 deaths, 14,551 confirmed cases, ratio is 2.1%.
However, one week ago we had 1312 cases, mortality becomes 23%.
3/ If we hypothesize that people died without going to the hospital, the ratio gets lower.
Also, lag estimates are averages; this skews the results.
Also, lag estimates are averages; this skews the results.
4/ (Am I doing any mistakes in the computation above? It seems making sense to me, and yet the numbers I see printed all around seem much lower. Please let me know if you have any feedback. Thanks.)
5/ Another challenge of lags is that some cities are way more infected than they know
Travel restriction measures should have been using Chinese data on contagions as a lagging indicator of the seriousness of the situation 7-10 days ago; not as a leading indicator for tomorrow's
Travel restriction measures should have been using Chinese data on contagions as a lagging indicator of the seriousness of the situation 7-10 days ago; not as a leading indicator for tomorrow's
6/ One of the most effective ways to slow down the contagion in already-infected cities is probably wearing masks.
City councils and citizens should be aware that their decision on whether to wear them should take into account the lag: they're seeing last week's contagion status
City councils and citizens should be aware that their decision on whether to wear them should take into account the lag: they're seeing last week's contagion status
7/ (The lag data used here, I found on multiple sources on the internet. It might or might not correspond to the real one; in any case, our samples are too small for now to say.
The point of this thread is not to produce exact numbers, but to raise awareness about a phenomenon.)
The point of this thread is not to produce exact numbers, but to raise awareness about a phenomenon.)
