August 14, 2020

expert reaction to preprint (not a published paper) looking at modelling city versus rural susceptibility to a second wave of COVID-19

A preprint from the University of Sussex, uploaded to medRxiv*, has modelled the susceptibility to a second wave of COVID-19 for rural populations versus city populations.

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“One thing that concerns me, about the models used in this research, is that they appear not to allow directly for transmission between people inside the metropolitan areas considered, and the rest of the country.  That might be a reasonable approximation if the metropolitan area has widely-drawn boundaries, with relatively sparsely population areas outside and not much commuting over the boundary – but I doubt if that applies to New York City or to London, even with much-reduced commuting because of lockdowns.  The fact is that the geographical nature of the four metro areas considered is very diverse.  The authors are not precise about which Madrid region they considered, but the only one in the data source they quote is the Comunidad de Madrid, which extends a long way from the city of Madrid itself; and Lombardy, while it does contain Milan, is a large and diverse region covering a lot of Northern Italy.  Some figures (from Wikipedia):

“Is exactly the same modelling approach valid for such a varied set of places, and in particular, do the detailed figures that the models give for London and for New York make good enough sense?”

Prof John Edmunds, Professor in the Centre for the Mathematical Modelling of Infectious Diseases, London School of Hygiene & Tropical Medicine, said:

“This study uses mathematical models to indirectly infer the level of immunity that there may be in major cities and the surrounding countries.  The authors suggest that significant numbers of people in London, Madrid and New York may be immune.  There is, however, a much more direct and reliable way to estimate the levels of immunity in populations; that is by taking a representative sample of the population and testing for specific antibodies against SARS-CoV-2 (the virus that causes COVID-19) in their blood.  Such studies have been done, including the ONS survey, the UKBiobank survey and the very large-scale REACT study.  These studies do indeed reveal that the level of immunity tends to be higher in London than elsewhere, however, they also show that the level of immunity – even in urban centres – remains relatively low.  For instance, the REACT survey¹ suggests that overall about 5-6% of adults in England have been infected with SARS-CoV-2, but rates of infection in London is about twice this (10-11%).  That is, the vast majority have not been exposed to this virus and remain susceptible, even in urban centres, contrary to what this study suggests.”

1 https://www.imperial.ac.uk/media/imperial-college/institute-of-global-health-innovation/Ward-et-al-120820.pdf

Prof Matt Keeling, Professor of Populations and Disease, University of Warwick, said:

“This is an interesting preprint, that puts forward the hypothesis already explored by both Gomes¹ and Gupta² that there is already far more immunity in the population than realised – but I am sceptical of the results.

“The proposition is supported by interesting data from several countries that show faster decline in major cities that were hit hard and early by COVID-19 compared to the rest of the country.  It is proposed that this is because the major cities have already depleted much of their susceptible pool, such that cases begin to decline earlier and quicker.

“The issue is whether this decline is due to a reduced level of susceptibles or a reduction in social mixing.  Figure 1b in this paper is put forward as evidence that London did not reduce its social mixing any faster than the rest of the country.  However, these are only proxy measures, and do not capture the true interaction between individuals that are likely to spread infection.  For example “transit stations” are presented as measures of social mixing, but this doesn’t account for mask use and only applies to particular sectors of the population.

“We therefore need to contrast this preprint’s predictions that 45% of Londoners and 19% of the UK could have been infected, with the recent data from the REACT study from Imperial³ which found 13% of Londoners and 6% of the UK were seropositive.

“While we would all like to believe that we are past the worst of this outbreak, the rising cases across many European countries would indicate that we are far from herd immunity and measures need to remain in force if we are to prevent a second wave even worse that the first.”

1 https://doi.org/10.1101/2020.07.23.20160762

2 https://doi.org/10.1101/2020.07.15.20154294

3 https://www.imperial.ac.uk/news/201893/largest-study-home-coronavirus-antibody-testing/

* Preprint (not a paper): ‘Impact of reduction of susceptibility to SARS-CoV-2 on epidemic dynamics in four early-seeded metropolitan regions’ by T. J. Barrett et al has been uploaded to medRxiv.  This work is not peer-reviewed.

https://www.medrxiv.org/content/10.1101/2020.07.28.20163154v1

All our previous output on this subject can be seen at this weblink:

www.sciencemediacentre.org/tag/covid-19

Declared interests

None received.