More

Background:
It is not the total number of contacts that is decisive for the risk of a virus spreading, but the contact index, at least seen statistically for the population. This is based on the so-called friendship paradox: a person with many contacts (purple in the picture) has a greater probability of being in contact with an infected person (red).


Therefore, the pure number of contacts underestimates the incidence of infection and must be corrected with the number of “supercontactors”. The blue curves show computer simulations in a graph of 1 million nodes (with SEIR model). One would naively expect the contact number and R to be linear. However, the true R is much higher. However, the expected linear behavior is obtained with the contact index corrected for the supercontacts.


This effect is extremely high in the contact networks we have determined! People have very different social ties and contacts (heterogeneous contact network).The number of people contacted on a day is very broad, i.e. there are many people who have very many contacts. That is why it is important to use the correction by the contact index and not the number of contacts.



The contact index actually shows a very high correlation with the case-based reproduction number. We were able to show this in a scientific study together with scientists from the Hasso Plattner Institute in Potsdam and the Charité in Berlin. So the case-based R-value can be predicted with high accuracy two to three weeks later.




Scientific original publication: (https://www.pnas.org/content/pnas/118/31/e2026731118.full.pdf)