“Without strong controls, the return to universities would cause a minimum of 50,000 deaths”
University and College Union, 30 August 2020
The University and College Union (UCU), a trade union for academic staff, tweeted a quote from a piece of research about the number of deaths from Covid-19 that the reopening of universities would cause. The UCU is campaigning to stop the reopening of university campuses, and its tweet was widely shared.
The tweet summarises research which models a scenario that will almost certainly not happen in the real world.
Where did “50,000 deaths” come from?
The UCU quoted a preprint by Professor Alan Dix, a computer scientist at Swansea University. (A preprint is an academic article that hasn’t yet been scrutinised by other experts and published in a journal.)
In the article, Professor Dix does indeed say in the first headline, “without strong controls, the return to universities would cause a minimum of 50,000 deaths” across the UK. Such “unrestrained” reopening would include the types of hygiene and social distancing measures already being planned, but not stronger controls such as students only being able to meet within small groups, or “bubbles”, an idea that the article considers later and which is similar to the scheme tested by some universities that returned earlier in the summer.
It also assumes no effect from other measures already in place, like the Test and Trace system or local lockdowns.
However, in the section where he calculates the 50,000 figure, Professor Dix explains that this is a “rough” estimate of “the potential impact of University returns” [our italics]. This means it is based on several assumptions that may not be true in the real world.
To begin with, Professor Dix explicitly assumes that “all efforts to limit spread within campuses fail”. Based on further assumptions about the rate that the virus would spread within universities, he therefore estimates that every student in the UK would catch the coronavirus during the autumn term.
In addition to this, the article assumes that these students will go on to spread the disease to non-students in the wider community. In his “most optimistic” scenario, Professor Dix assumes a reproduction rate (“R”) in the community of 0.7, meaning that each infected person directly infects an average of 0.7 further people, who in turn also infect an average of 0.7 each, and so on. This would eventually mean that each infected student would lead to an average of 2.3 further infections among non-students.
The article calculates that this means “between 5% and 6% of the general population will become infected due to the university impact”. It says that this is the same as were infected during the first wave of the pandemic in Britain, which appears to have caused about 50,000 deaths, based on the number of “excess deaths” reported at the beginning of May. (The number of excess deaths is a measure of all deaths, from any cause, above what might have been expected.) For this reason, Professor Dix’s rough estimate in this scenario is that the UK would experience the same number of deaths again, as a direct result of universities reopening.
But will this happen?
We contacted Professor Dix, who told us that in his opinion, “all credible scenarios lead to massive figures [of Covid-19 deaths]”. However, he also told us that “this [50,000] death toll will not actually happen in its entirety, because once it became apparent it would trigger local or national lockdown, and/or the universities themselves being closed down.”
We also spoke to Dr Kit Yates from the University of Bath, who told us that in his view there are some mistakes in Professor Dix’s model. “I'm not completely damning the work,” Dr Yates says, “because I think the return of University students will lead to an increase in spread and could have a negative impact, I'm just not convinced it will be quite on the scale that is predicted.”
In short, by tweeting the quote without context, the UCU might have led people to believe that reopening universities was likely to cause at least 50,000 people to die of Covid-19. In fact, this figure comes from a model that is open to question and which makes predictions that will almost certainly not play out in the real world.