13 November 2020

New transmission model accounts for superspreaders


Researchers at the University of Copenhagen are the first in the world to create a mathematical model that takes superspreaders into account when predicting the spread of COVID. The model demonstrates that using the so-called contact number is inadequate for understanding how COVID spreads, as not everyone is equally infectious. Using the model, researchers are also able to prove that isolating ourselves in social bubbles is unnecessary. Being in social clusters, where people have contacts beyond their clusters as well, is enough to inhibit superspreaders and reduce infections.


COVID-19 is a disease in which about 10 percent of the population has the infectious potential to cause 80 percent of transmission, as some people — referred to as superspreaders — have been shown to be far more infectious than others. Observational studies have already pointed this out. Now, researchers from the University of Copenhagen’s Niels Bohr Institute and Roskilde University (RUC) are equipped with a new mathematical model, one that is the first to take the superspreader effect into account when predicting outbreaks of COVID. (https://doi.org/10.1101/2020.09.15.20195008)

"That not everyone infects equally is incredibly important when you want to control a pandemic. Previous models calculated under the assumption that all humans were similar. Doing so works well with influenza, but not with COVID-19," explains PhD Bjarke Frost Nielsen, who together with Professor Kim Sneppen of the Niels Bohr Institute and Lone Simonsen from RUC, is behind the research project. The research results have been published provisionally as preprints to make the latest research available during the pandemic. They will be published in a scientific journal at a later date.

Contact numbers cannot stand alone

The models that have been used up until now to predict infections have generally overlooked the importance of superspreaders. This is because the models are of the so-called well-mixed type, which cannot account for differences at the individual level. While this works for many diseases, there are major differences in transmission rates among individuals with COVID, which means that more sophisticated mathematical models are required.

The new model allows for far greater detail, making it possible to take individual variations in contagion into account, including superspreading. The model incorporates what we know about the distribution of contagion in COVID-19 into a so-called agent-based model where individuals — as opposed to entire populations — can be followed individually.

 The new knowledge means that the contact number, or reproduction number, known as the R number — which describes how many people the average corona patient infects — is an insufficient parameter for understanding the evolution of COVID infection rates in Denmark.

"The R number only represents an average, but one also needs to look at the spreading in terms of how broadly individuals infect. This spreading has decisive consequences for how best to manage COVID transmission,” explains Bjarke Frost Nielsen.

Social clusters and size limits on social gatherings are the strongest cards

The new mathematical model demonstrates that COVID-19's tendency towards superspreading is actually good news.

"COVID’s Achilles heel is that it is a superspreader disease. This makes it much more susceptible to restrictions on gathering sizes and personal contact networks", says Lone Simonsen.

The model suggests that it is safe to see about 10 people regularly, as long as large public gatherings are also avoided. As such, the research supports the correctness of actions taken by the Danish government. However, the 10 people one sees don't all need to be the same people. The researchers recommend that we talk about social clusters instead:

"Our latest research shows that social clusters are one of our strongest cards in the fight against COVID. But there is no need for social bubbles.  Social bubbles are about trying to isolate ourselves from the rest of society. Doing so isn’t necessary. Social clusters, on the other hand, are simply about avoiding seeing too many people who never have contact with each other", explains PhD Bjarke Frost Nielsen.

The research results have been published provisionally, as preprints, in order to make the latest research available during the pandemic. They will be published in a scientific journal at a later date. The research results are on their way to being peer-reviewed and published in a scientific journal. Link to the scientific article in preprint.