June 5, 2020

expert reaction to latest COVID-19 nowcast and forecast by the MRC Biostatistics Unit at the University of Cambridge and PHE

The MRC Biostatistics Unit at the University of Cambridge, working in partnership with Public Health England (PHE), have published an update of their COVID-19 nowcast and forecast figures.

Prof Sheila Bird, former Programme Leader, MRC Biostatistics Unit, University of Cambridge, who was not involved in this work, said:

“Today’s report, led by Professor Daniela de Angelis and Dr Paul Birrell, is the regional update that everyone has been eager to have. But the findings are the opposite of reassuring.

“First, in no region of England is the effective reproduction number assuredly below 1 with regions’ median estimates ranging from 0.89 to 1.01. No wonder members of SAGE are worried.

“Secondly, the estimated daily number of new infections is 16,700 (uncertainty interval: 11,000 to 25,000). That is three times this morning’s estimate from the ONS Infection Survey of 5,600 new infections daily for persons who reside in households (95% CI: 3,700 to 7,900).

“The Cambridge/PHE team has taken account of seroprevalence estimates provided by the National Blood Transfusion Service, about which information has been released in this afternoon’s deluge of SAGE papers. What assumptions have been made about the persistence of antibodies in incorporating these seroprevalence and, indeed, about the specificity and sensitivity of the test-deployed in their assessment is not immediately apparent but of course matters.

“The Cambridge/PHE team do not appear to have made use of earlier estimates from the ONS Infection Survey. Also, whereas the team’s previous report indicated that they were working to incorporate information on all COVID-mention deaths registered by ONS (which requires nowcasting of those deaths), the update does not make clear if this objective was achieved.

“Monthly updates of regions’ effective reproduction number are essential, if uncomfortable, reading. Appraisal by SPI-Modelling will be needed to interdigitate different estimates of apparently similar, but actually somewhat differently-defined, quantities such as ‘daily number of new infections’.

“Whichever estimator for daily new infections you choose, monitor how it changes over time and how your chosen estimator compares with the daily number of COVID-19 swab-test positives, currently less than 2000. Mind that gap – new versus diagnosed – because it’s the gap that matters.”

https://www.mrc-bsu.cam.ac.uk/now-casting/

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

www.sciencemediacentre.org/tag/covid-19

Declared interests

Prof Sheila Bird: “The Cambridge/PHE report has been led by former colleagues at MRC Biostatistics Unit but I was not involved in the work.”