This paper argues that having a mutually-consistent and reasonably small network of contacts can help in controlling coronavirus spread…for example, if a group of 7 people work together and also socialize together, they are better-off from a potential infection standpoint than if individuals in the group are socializing with different, and frequently changing, sets of people.
Somewhat related: the Federal Aviation Administration is taking steps to limit the spread of coronavirus in air traffic control facilities:
Each air traffic control facility is establishing separate teams of controllers that will stay together throughout the duty week. Each crew will contain the same employees, limiting the possibility of cross-exposure to COVID-19 that would come through normal shift rotations. If a person on one team gets sick, the only people who would be exposed are the other people on that team.
So, presumably, if one member of a team gets sick, all the team members would go home until they can get tested and found coronavirus free, and a new team will be swapped in to support operational needs. Not sure how large these teams are: in a control tower for a medium-sized airport, a team might consist of all the people working on a particular shift, but for a large facility like Potomac Approach or Kansas City Center, I imagine that the teams must comprise only subsets of the total workforce; probably people who work in close proximity to one another.