Risk calculation when exposed to multiple infectors both for < 15 min.

[r-r-liu]

In his latest podcast (47, https://www.ndr.de/nachrichten/info/podcast4684.html) C. Drosten seems to be saying that there is evidence of people being infected with different strains of the virus simultaneously, leading him to suspect that it is possible to be infected by more than one person. I'm not sure whether he means that the amount of viral matter of each strain was sufficient to infect the subject, or whether it was the sum. (I hope I understood that, I'm not a biologist, virologist, epidemiologist, etc.) Does the risk calculation take that into account? If I understand the documentation on the risk calculation, that is now being performed in such a way that exposure per source is not being calculated, just total exposure (to however many sources) for a sufficiently long time. Is this so?

[timokau]

The way I understood it, his assumption is that the vast majority of infections still come from a single source. Its just that the "channel" is wide enough that multiple virus strains might be transmitted at the same time.

He hypothesizes that the amount of virus that is initially transmitted may influence the severity of the illness.

[r-r-liu]

Timo, OK, thanks. Regards, Richard

…

On 10 Jun 2020, at 23:33 , Timo Kaufmann ***@***.***> wrote: The way I understood it, his assumption is that the vast majority of infections still come from a single source. Its just that the "channel" is wide enough that multiple virus strains might be transmitted at the same time. He hypothesizes that the amount of virus that is initially transmitted may influence the severity of the illness. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[keugens]

I think this question is very valid.

Consider a meeting with an asymptomatic group of persons. A group, which lives in the same household and has probably the same virus regarding its mutation and development. From a physical point of view, I would say: yes, the virus flow will sum up. If an infection starts after 15 min with one person, it will start approximately after 8 min with two persons.

From an epidemiological point of view, one could answer: the initial dose is very small, therefore the accumulation from multiple sources and over time does not really matter. The 15 min threshold is applied only with respect to the probability to receive this small initial dose.

That might be true. But again, from my humble point of view, I would answer: multiple sources will increase this probability as well.

Therefore, for me, this seems to be the way to go, as long as there are no studies to prove the opposite.

[r-r-liu]

Remember, there's a cost for false positives, i.e., voluntarily going into quarantine. Even if one's salary is paid during that time, one's belief in the app should not be taken for granted. Just dead reckoning (𝛑 × 👍), two false positives is probably the most anybody can be expected to accept before ignoring the next warning.

I understand that some parameter for computing the exposure risk are passed to the app. If that so, are they such that the case of multiple sources could be covered in by simply adapting one or more of them?

[keugens]

Theoretically yes. For multiple sources - all lower than 2 meters in distance - you could calculate a reduced duration threshold. But as mentioned in the LASCEC report, the knowledge of multiple sources is a bad thing due to privacy.

[r-r-liu]

I return to this point having read (1) the document on exposure score calculation (several times) and (2) an article in today's (2020-09-05) Basler Zeitung, "Nähe zu Testcenter führt zu Covid-Fehlalarmen bei Pflegekräften".

Questions about my understanding of the exposure score calculation:

Are these observations of mine correct?

• The document describes a calculation that seems not to distinguish between various sources (i. e., app users who have tested positive) of the exposure keys.

• In the end, it seems to me, each attenuation range bucket contains a single number representing the total duration of exposure to all users who submitted keys for the day in question, and so, if I am not mistaken, the app warns if the sum of the attenuation-weighted exposures exceeds 15 minutes per day.

My reason for concern:

The article in the BaZ describes a situation where a Covid-19 test center of BL has been opened in a building that houses classrooms for students studying to become healthcare professionals. The students walk past lines of people waiting to be tested several times a day, may even ride public transportation with them to arrive at the same building, and have been receiving what they, resp. the health authorities of BL, consider to be false warnings. So the students turn off exposure logging in those situations because they clearly have not been exposed to any single person waiting to be tested for 15 minutes or more. If my understanding of the exposure score calculation is correct, the app is behaving as designed and has very good reason for doing so. For, although the app doesn't know this, the students whose classes take place in this building are specializing in the care of the elderly, and after classes many of them have contact with elderly patients in homes or hospitals!

Again, if my understanding of the exposure score calculation is correct, what the app is doing must be communicated to the public so that users can make an informed choice, when to turn off exposure logging and when not to.

[r-r-liu]

I have found the answer my question here: https://github.com/admin-ch/PT-System-Documents/blob/master/SwissCovid-ExposureScore.pdf . Assuming that document in admin-ch is still current, the answer to my question is yes. The document in admin-ch is very explicit on this point:

"In Switzerland, an exposure notification should be shown to a person if that person has been exposed to COVID-positive individual(s) for 15 minutes or more during one day (​Ordinance of 6/24/20 of the Swiss Federal Council). The Annex of the Ordinance specifies a threshold distance of 1.5m, which is lower than the 2m threshold established by the ​ECDC." (pg. 7)

So is the text to which "Ordinance" links:

"si la comparaison révèle qu’un téléphone portable s’est rapproché à moins de deux mètres d'au moins un participant infecté et que la durée totale de cette proximité atteint ou dépasse les quinze minutes au cours de la même jour- née, l'application envoie une information." (pg. 4).