April 8, 2022

expert comment about SARS-CoV-2 surveillance, including REACT-1 and the ONS Infection Survey

A comment from Professor Sheila Bird on COVID-19 surveillance in case useful.

Prof Sheila Bird, Formerly Programme Leader, MRC Biostatistics Unit, University of Cambridge, said:

“Marvellously, 2 million persons in England have been participants in the 19 rounds of REACT-1.  Yet, the public’s participation-rate [15%] does not really do justice to the insightful analyses that REACT-1 has offered us.  Each round showed us both unweighted and re-weighted estimates as analyses were adjusted to compensate for under-representation.

“‘REACT going out on a high’ is how one of the team described to me their final report earlier this week which underscored the worrying extent of transmission of Omicron BA.2 in late March 2022: in those aged 55+ years: yet, I had just returned from a not-so-youthful scientific meeting in London where only 10% of the audience was masked!

“Digression on cost-efficiency in science: Turning briefly from public health to the public purse, l am cannily aware that “mony a mickle mak’s a muckle” (small contributions add up to a great deal).  REACT-1 did not run in Scotland.  By all accounts, REACT’s annual cost is about one-tenth the cost of ONS’s Community Infection Survey (ONS-CIS).  And Scotland’s ONS-CIS sample is only about one-tenth of the CIS-England.

“Does this mean that Scotland could instead consider investing its surveillance revenue in REACT-1, at say 80,000 participants per round in Scotland, giving Scotland substantially greater precision and in-depth analysis for its investment?

“Back to public health: Scotland’s COVID-mention deaths by week of occurrence in mid-March 2022 (likely 3-week total: around 500, see Table 1) already exceed Scotland’s 2-week Omicron BA.1 peak of 278 in mid-January 2022.  England’s COVID-mention deaths (mainly Omicron BA.1) were at their nadir in the occurrence-week ending 4^th March since when they have increased steadily (see Table 2), an increase that REACT implies will continue well into April 2022.  Remember that epidemiological week is defined slightly differently for England & Wales, see Table 2.

“Table 3, based on national proportions testing SARS-CoV-2 positive in the ONS Community Infection Survey, not only supports REACT’s alert for England but shows that Omicron BA.2 has hit harder and earlier in Scotland when compared with Omicron BA.1 in Scotland [over 50% harder] or versus Omicron BA.2 in England & Wales [three weeks earlier], see Table 3.

“Finally, notice in Table 3 that the width of Scotland’s 95% confidence intervals is about three times greater than for England.  REACT for Scotland could probably equal England’s precision on a cost-neutral basis.”

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

www.sciencemediacentre.org/tag/covid-19

Declared interests

None received.