January 19, 2021

expert reaction to ONS stats on deaths registered weekly in England and Wales, provisional: week ending 8 January 2021

The Office for National Statistics (ONS) have released provisional counts of the number of deaths registered in England and Wales in the week ending 8 January 2021.

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

“The weekly bulletin on death registrations from ONS takes the numbers forward to the week ending 8 January 2021.  Some of the numbers continue to be difficult to interpret, as they have been for the last couple of weeks, because of registration delays over the Christmas holidays, and because the comparisons with previous years are more difficult because of when Bank Holidays fall in relation to the numbered weeks used for comparison.  Of course, the Christmas holidays were over by the latest week (which started on 2 January), but registration delays will still matter.  They will tend to work in the opposite direction compared to weeks where there is a holiday, with some registrations being ‘pushed back’ to the week ending 8 January if they did not occur in the holiday weeks.  So ONS are right to warn that the very sharp increase in numbers of deaths in the latest week compared to the previous two weeks may well, to a considerable extent, be due to these delays connected with our human calendar customs.

“But there are other ways in which we can get a check on whether deaths really did go up in this first week of 2021 compared to previous weeks.  One is that ONS also publish numbers of deaths in different weeks according to when the death actually occurred, rather than when it was registered.  Those numbers are affected by registration delays too, but in a different way.  The numbers by occurrence date for the latest week include deaths that occurred up to 8 January, and that were registered up to 16 January.  Some deaths in the week ending 8 January will still not have been registered by the 16^th.  So what tends to happen, at times when numbers of deaths aren’t changing fast, is that the figures by date of occurrence for the latest week always look low compared to previous weeks, when they are first published, but they are revised upwards later as more registrations come in.  (That’s one important reason why ONS call these figures ‘provisional’.)  But, even allowing for this effect of late registration on data by occurrence date, the number of deaths involving Covid-19 in the week ending 8 January was 7% higher than the previous week, and indeed higher than the figure in every week since the start of May 2020.  The number of deaths from any cause, by date of occurrence, was also higher in the week ending 8 January than the previous week, though by a smaller margin (2%), and again is the highest is has been since the start of May 2020.  Those increases will look considerably larger when the numbers of deaths by date of occurrence for the week ending 8 January are revised upwards as more registrations come in.  This is grim, really grim.

“The ONS data, by date of registration, indicate that 34% of the deaths registered in the week ending 8 January mentioned Covid-19 on the death certificate, and indeed that 30% of the deaths registered that week had Covid-19 coded as their underlying cause (so not people who died of something else but happened to have Covid-19, which may have hastened their death but didn’t directly cause it).  So it’s absolutely clear that the untypically large numbers of deaths in that week are because of the pandemic.  Measured by date of occurrence, 36% of the deaths that week involved Covid-19.  (I have no data yet on deaths that occurred that week and had Covid-19 recorded as the underlying cause.)  Again because of issues about registrations over the Christmas holidays, it’s harder to have a good measure of the excess deaths above the expected number in the latest week, though all indications are still that the excess number is truly pretty high.  And generally the grim picture applies across all the English regions and Wales.  Numbers of deaths in the week ending 8 January are above the average of previous years in all the regions and in Wales, and above the figure for the previous weeks.

“Depressingly, none of this is at all surprising, given the news in recent weeks about the progress of the pandemic, in terms of confirmed cases, hospitalisations, and deaths.  Generally across England and Wales, the peak number of confirmed cases was in the last week of the Old Year, though this varied a bit between regions.  But, in the sad event that someone dies of Covid-19, they would typically have been infected maybe three weeks before.  So the deaths reported in this bulletin would generally have involved people who were infected when cases were still rising.  There are some indications that the daily numbers of deaths recorded on the Government dashboard are now levelling off, or indeed just starting to fall – but that couldn’t show up in these registration figures up to 8 January, and they may well not fall in next week’s registration figures either.  And the current level of Covid-19 deaths is so high that it has to fall a very long way indeed before we’re out of the current serious crisis.”

Prof Sir David Spiegelhalter, Chair, Winton Centre for Risk and Evidence Communication, University of Cambridge, said:

“Compared to the five-year average, since the start of September there have now been over 26,000 excess deaths registered in England and Wales.  Over this period around 31,000 deaths had Covid as their primary cause, so this means there have been around 5,000 fewer deaths from ‘normal’ causes.  These are not wrongly labelled as Covid, so where have these deaths gone?

“First, it’s been a mild winter so far, and the anti-Covid measures have been enough to stop flu in its tracks.  There will also have been many older vulnerable people who were among the 59,000 excess deaths between March and June, who would ordinarily have survived until later in the year: in contrast to what was sometimes claimed at the time, less than 10% of those spring deaths would have died anyway later in the year.

“One reassuring interpretation is that the lockdown measures and disruption of the health services have not yet caused a notable increase in non-Covid fatalities.”

https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/deathsregisteredweeklyinenglandandwalesprovisional/weekending8january2021

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

www.sciencemediacentre.org/tag/covid-19

Declared interests

Prof Kevin McConway: “I am a Trustee of the SMC and a member of the Advisory Committee, but my quote above is in my capacity as a professional statistician.”

Prof Sir David Spiegelhalter: “DJS is a paid non-executive director of the UK Statistics Authority, that oversees the work of the ONS.”