Is it really fair to blame universities for the second wave?

You may well have never heard of Watton.

It’s a small market town in the Breckland district of Norfolk. Early last week it had the dubious distinction of having the one of the highest Covid-19 case rates in England. An outbreak centred around the Cranswick Food factory resulted in mass testing and the discovery of 140 cases on 26 October.

There’s no indication that there is a rise in cases in wider Breckland. Watton is an example of what is supposed to be happening everywhere – an outbreak (defined as more than one case in a single setting) is reported, and local public health teams carry out mass testing to identify other cases before establishing safety measures that could range from physical changes to the work environment to temporarily closing the setting (in this case the factory remains open, though it has temporarily ceased exporting to China).

There are always a handful of stories like this – they make the local news or get drawn into wider arguments about working conditions. But the majority of Covid-19 cases in the UK are not linked to outbreaks.

How outbreaks work

University halls of residence are probably the best reported instances of multiple outbreaks – we saw numbers of cases rise quickly and dwarf those reported in other areas. We’ve read so much about Covid on campus, and the numbers looks so terrifying, that you would be forgiven for assuming that they represent a significant impact on the total number of cases in England.

Here I’ve plotted the historic seven day rolling rates of cases for every MSOA (small area) in England. This is the best data we have on how much Covid was in circulation in a given area at a given time. And you can see where both the volume and rate of second wave cases has come from – and at a national level it is not the areas full of students in halls of residence. The overall growth comes from a steady rise in places with few or no students. I’ve used the colours to show areas with lots of students in term-time residence (over 1,000 – so likely to have one or more halls of residence in the area) in dark blue and those with a still substantial number of students (250 to 1000) in orange.

[Full screen]

Late September and early October looked scary, as university terms started across the country. But as we move out of October the peaks (primarily around halls of residence) drop while the background rate continues to rise – and in cities in the south of England the same pattern will become apparent. And as we’ve seen, it is the background that drives the rise in overall volume – there may still be peaks around some universities, but these will also fall away quickly.

Alternate reality

What would have happened if students had not returned to campus this autumn? Clearly the peaks in student areas would not have happened – cases around individuals who arrived on campus carrying the infection would have been scattered around the many areas where young people are domiciled. Others would still have been infected, but outbreaks would not have been identified, and the many asymptomatic cases that drive the hall of residence peaks would not have been detected. The wider rise in cases from early September would have still happened – perhaps, with less rigorous testing and isolation among the young, at a faster rates.

The longer term impacts of Covid-19 on young people is still largely unknown. We don’t know the prevalence of “long Covid” or the risk factors associated with it. It is a risk to young people and young people should not have been exposed to it in the way that the mass migration at the start of term led to. As things look currently, the risk in students returning home is probably lower than the risk of moving about a million other people around the country. Still not something you’d do lightly, but a lot better than it would have been running another mass migration even four weeks ago. January may well be another matter, of course.

It could be argued that the early student peaks offered an early warning of the wider second wave. The government did not spot this, but neither did anyone else.

Bonus data

I’ve put together a location based dashboard for you to explore this data further. The colours of areas relates to numbers of students in an area, the text on the map to the number of cases. The filters control both graphs, and you can click on a mark of interest on the graph to see the location of a small area, and use the date slider to track changes over time.

[full screen]