Cast your mind back to the early October campus covid case peaks, and you might recall a rash of press stories about misattributed cases.
The tale went that because many students were still registered with a GP at their home address, positive tests taken while at university were erroneously being added to area totals away from the campus and environs. It was very much a precision rather than a direction of travel issue – we already knew about the huge peaks of infection around universities, and whether or not the number was larger or smaller didn’t change what we needed to be doing (testing, tracking, tracing, isolating) or what we knew about the virus.
In search of accuracy
But it rankled with many that we didn’t have accurate numbers of positive tests for university areas. And yesterday saw an attempt – in England, at least – to fix this issue. Location is now derived from the area where the test was taken, rather than the address registered with the NHS.
We have details around how many cases were involved – between 1 September and 12 October the issue involved 12 per cent of the total cases in England, with the last week in October seeing this number drop to 6.6 per cent (of, of course a much larger number). It’s worth keeping a model of the shape of the second wave in mind – cases started rising (especially in the North of England) from early September, there was a peak in some student areas around the start of October but the underlying rise continued even as this peak dissipated.
What does this mean
Could corrected locations have changed the shape of the response to the second wave? The government reassures us that:
The conclusion is that recording location based on the NHS database has not affected any such decisions [on local or national restrictions]”
This issue would not have affected contact tracing by NHS Test and Trace”
Now at this stage I am, as I am sure all readers are, inclined to take government pronouncements about the course of the pandemic with more salt than I do on my chips. But the data does appear to back this up.
Fire up the tableau!
Inspired by the consistently superb visualisation work of @RP131 on twitter, I’ve compared the data released yesterday (which includes all the historic changes) with the data released on Sunday (which is the last release of the data by registered NHS address). The data is available in 7 day periods which don’t quite match up – I’ve done an approximation to link the data, but you should see these values as indicative only.
Here’s a plot (from left to right) of the old figures, the new figures, and the difference between the two. You can filter by local authority area and then, if you like, by individual MSOA. I’ve used colour to show the number of students living in each MSOA – anything with even a tinge of blue has more than 1,000 students (so probably at least a decent-sized hall of residence).
First thing you should spot (the plot defaults to all areas in England) is the shape of the two Covid-19 waves, and the smaller blue peak relating to student area cases underneath wave two. The second wave clearly started before students returned, and there has never been a week where students have represented all or most of the cases in England.
We can see that the change from registered to supplied addresses has had a significant effect on student areas (third graph) with a consequent negative impact on non-student areas – which may represent home domiciles. But other areas have had cases added too – one could imagine barracks, care homes, and residential schools having an impact here. But don’t forget we have exposed a wider issue of poor quality address data in the NHS database – this needs fixing, and data-sharing between local authorities, HMRC, and the NHS may provide the answer.
Newcastle and Leeds show a notable impact with a large number of cases in student areas in the early part of the second wave. But it is equally notable how quickly student area cases dropped following the early peak, and how consequent growth has come from non-student areas. As I’ve examined on the site before, this suggests cases are not spreading as rapidly within student areas as elsewhere – better isolation practices, greater testing availability (including asymptomatic testing), and university support will all have played a part in driving R below one in these areas. And if R is below one in a small area it is unlikely to be spreading to a neighbouring area (there’s a map if you want to look at geography and scroll through the weeks).
There is still a pervading myth that the decision to bring university students to campus started the second wave – this is emphatically not true. We were clearly going to have a second wave of Covid-19 whatever we did – bringing students to campus to experience this must rank alongside “eat out to help out” and “back to the office” as one of the sillier ideas of the summer, but unlike the latter it was more a matter of a terrible experience for students and staff than a public health disaster. It’s particularly troubling to see university staff complicit in spreading the myth that everything is the fault of the student.