Matt Hancock has announced an urgent review into how Public Health England (PHE) counts Covid-19 deaths after discovering what appeared to be a serious issue in how rates are calculated.
Following the health secretary’s move on Friday, Yoon K Loke and Carl Heneghan, of the Centre for Evidence-Based Medicine at Oxford University, wrote in a blogpost: “It seems that PHE regularly looks for people on the NHS database who have ever tested positive, and simply checks to see if they are still alive or not. PHE does not appear to consider how long ago the Covid test result was, nor whether the person has been successfully treated in hospital and discharged to the community.”
A Department of Health and Social Care source summed this up as: “You could have been tested positive in February, have no symptoms, then be hit by a bus in July and you’d be recorded as a Covid death.”
When the Guardian put this to a source at PHE, they said that such a scenario would “technically” be counted as a coronavirus death, “though the numbers where that situation would apply are likely to be very small”. PHE says it calculates deaths in this way because, in most circumstances, it cannot dismiss the possibility that Covid-19 could have played a role in the death.
But does this mean PHE’s figures are much higher than the actual virus death rate? Not quite.
The controversy is confined to England (devolution means deaths are counted differently around the UK) and is best demonstrated by the different ways in which PHE and the Office for National Statistics (ONS) count Covid-19 deaths. The former reports deaths as they are reported to it; the latter’s count takes longer but is more comprehensive.
At the beginning of the pandemic, the PHE figures only included deaths that occurred in hospitals, later expanding to include deaths in care settings. Critically, however, its figures only include deaths where the patient has first tested positive for coronavirus, and early in the pandemic there was a lack of testing.
The ONS, by contrast, includes all deaths where Covid-19 appears on the death certificate.
Up to the end of June, the ONS figures were consistently higher than those released by PHE. That trend has reversed in recent weeks where the PHE figures have been higher than the ONS count except at the weekends, where reporting of deaths is lower but people are still just as likely to actually die on a Saturday or Sunday.
As things stand, the official PHE figures are still a significant undercount of the true death toll, which is more accurately reflected in the ONS data.
Figures up to and including 3 July show that 48,154 Covid-19 deaths had occurred in England, a figure which will increase as more deaths are registered.
The total announced by the government to the same date stood at 39,626.
However there appears to be a general agreement that, as time goes on, the PHE approach could become problematic. Indeed, Dr Susan Hopkins, PHE’s incident director, was quoted in a tweet on Friday saying: “Now is the right time to review how deaths are calculated”.
Prof Sir David Spiegelhalter, chair of the Winton Centre for Risk and Evidence Communication at the University of Cambridge, said: “There would not have been a big impact until recently, when there has been time for people who have tested positive to recover and then die of something unrelated.”
Speaking at a meeting of the Independent Sage committee, Prof Christina Pagel, a professor of operational research at UCL, said: “If someone [got Covid-19] in mid-March, recovered early April, the chances of them then dying from something completely different in the last couple of months is quite low. So I don’t think it has caused a massive distortion.”
Oxford’s Loke said that, at present, the overall death toll is unlikely to be affected since early in the outbreak the PHE approach underestimated deaths from Covid-19 due to a lack of testing. “The true Covid death toll probably is close to what PHE and ONS are reporting, but the time pattern with PHE is wrong,” he said.