In one well-studied example, at a choir practice with 61 attendees that took place one evening in March, in Skagit County, Washington State, one person infected an estimated 52 others. Doctors followed up with the close contacts of every secondary case, about three or four each, and could find only 10 further infections, says Lea Hamner, a public health official in Skagit County. It is as if something qualitatively different were going on that night.
What might that be? Transmission clusters have been seen with other diseases, including HIV, TB and typhoid, with a famous superspreader being a New York cook in the early 20th century who came to be known as Typhoid Mary. In these cases, it seemed there was something biological that made the person more likely to pass on their germs, probably a heavy microbe burden.
With coronavirus, virus burden may well play a role, but this hasn’t been investigated and we have no easy way to start doing so, says Benjamin Cowling at the University of Hong Kong in China. “If we measure viral load in saliva, that’s not the same as how much virus they’re breathing out. You would have to do some kind of air sampling.”
But as well as biology, the circumstances of the spreading event also seems to be important, and some common themes have emerged. Cowling’s group carried out contact tracing of the first 1037 coronavirus cases in Hong Kong. They found a somewhat larger K value than the previous estimate, of 0.45, but that still means just 20 per cent of infected individuals caused 80 per cent of locally acquired cases.
The team found that superspreading events tended to happen in indoor spaces, with people in close proximity. Social occasions led to more clusters than exposure in the workplace or home – mass transmissions occurred at weddings, temples, bars and karaoke parties, for instance. The risk seems to be higher if people are raising their voices in some way, such as singing or shouting. “It’s the volume of air that comes out of your lungs,” says Cowling.