This is a thread that attempts some back-of-the-envelope calculations to assess how the ban on flights from Europe to the USA could affect the number of #COVID19 cases in the US. These sorts of calculations are often called Fermi calculations (en.wikipedia.org/wiki/Fermi_pro…). 1/14
I'm only attempting to do a broad comparison of the number of new infections introduced from Europe daily to the number of new infections generated endemically by the growing US outbreak. 2/14
First, new introductions from Europe. One important data point is the @MRC_Outbreak report that on Feb 1 flights from Wuhan had about 1% prevalence (imperial.ac.uk/mrc-global-inf…). This was when there were 14k total cases reported. 3/14
I'm going to multiply confirmed cases by 5 to get an estimate of current prevalence. Note that this is approximately what we'd expect from WA state right now with 338 reported cases and 1000 to 2000 active infections. 4/14
We have close to 20k reported cases in Europe (who.int/emergencies/di…), primarily in Italy, Spain, Germany, France and Switzerland, and so an extremely rough estimate of 100k active infections. 5/14
The population of these countries is 260 million. This would give a current prevalence in these countries of 1 infection in 2600 people. 6/14
There are roughly 250k passengers per day from Europe to the US (transportation.gov/office-policy/…). If we assume that sick people travel just as often and that air travel would be at normal levels without ban, we get roughly 100 introductions from Europe to the US every day. 7/14
Next, new infections from within the USA. There are 1289 confirmed cases today (nytimes.com/interactive/20…) and so I estimate extremely roughly 6500 active infections in the US at this moment. 8/14
If we expect the number of infections to double in 7 days, that means that our current rough estimate of 6500 cases will be 13,000 cases 7 days from now. This means that at the moment, there are 13000/7 = ~1800 new infections per day. 9/14
There are many assumptions here (particularly the ratio of 5 to 1 active infections to confirmed cases), but the key result which should be fairly robust to these assumptions is the ratio of 1800 endemic infections vs 100 travel introductions from Europe. 10/14
This is a ratio of 18 to 1 for the number of new infections generated within the USA vs infections brought in from Europe. 11/14
I've tried to be conservative in this estimate, sources of error (lack of testing in US vs Europe, current travel behavior) would throw things in the direction of higher ratio of endemic infections rather than lower. 12/14
I believe the focus needs to be on testing and case finding in the US to slow transmission here. Introductions through air travel are likely to have a more minor impact at this point. 13/14
(Revised tweet #13. I felt more exact language was necessary.)
Follow up: Not all cases are the same. New introduction events into areas without endemic US circulation is an issue. Some fraction of new endemic infections could introduce the virus to a new geography, but not all will, while European introductions may be more likely to do so.
