select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to latest ONS stats on deaths registered weekly in England and Wales, provisional: week ending 7 January 2022

The Office for National Statistics (ONS) have released provisional counts of the number of deaths registered in England and Wales in the week ending 7 January 2022.


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

“While it’s not encouraging that death registrations involving Covid-19 were higher in the week ending 7 January than the previous week, it’s not surprising either.  Deaths on the Government dashboard in England and Wales were rising from roughly the last week of December until at least 7 January, in the figures classified by the date the person died and not the date that the death was reported (because there were many reporting delays over the Christmas period).  There’s some indication from the dashboard data that the increase might have flattened off more recently, but that’s not clear yet, and if it has happened, it’s too late for it to have showed up in the newest ONS death registration data.

“The increase in registrations involving Covid-19 in the latest week compared to the week before is larger than the dashboard increases, but that’s very likely because of registration delays because of the bank holidays.  I don’t think it makes much sense to look at short-term trends in death registrations in this period, for Covid-related deaths, all-cause deaths, or excess deaths.  As ONS rightly point out in their bulletin today, “We often see high death registrations in the first two weeks of January when registration services are back in office and dealing with any backlog from the bank holiday period because of the Christmas period.”  And indeed, in the most recent week, ending 7 January, not only was there New Year’s Day itself, but the associated bank holiday was on 3 January because the 1st was a Saturday.  The registration figures might not entirely be back to normal next week.

“Some have asked whether Covid deaths are roughly replacing deaths from flu and other ‘winter’ diseases – but I’d say that generally it’s difficult to tell how far that makes sense.  You certainly can’t say from this ONS release – apart from the Bank Holiday disruption to the numbers, the bulletin doesn’t give much in the way of relevant information.  There are figures for deaths involving all respiratory diseases, and for influenza and pneumonia.  One might think that the flu and pneumonia numbers in this bulletin are helpful for comparisons with flu, but in fact they are not very helpful, because of the way flu deaths are coded.  Few deaths are coded as being due to influenza – most of the ‘flu and pneumonia’ numbers are for pneumonias.  Agencies like UKHSA (and before them PHE, and the equivalents elsewhere in the UK) estimate flu deaths using statistical models based on a range of other data.  UKHSA issue regular weekly surveillance reports1 on flu and Covid-19, and they are indicating low levels of flu infection so far this winter.  But how far that is due to the restrictions as in England’s Plan B, or to people taking more care with contact, masks and so on than they would have done before Covid-19, isn’t really clear yet, not to me anyway. Certainly the lockdown last winter kept flu and other respiratory infections at a very low level indeed.  But I haven’t (yet) seen any estimates for current levels of flu deaths.

“Some are interested in whether this data tells us anything about people dying with rather than from Covid and whether this is different with omicron than with previous variants – for the most recent week (ending 7 January), 77% of the death registrations, where Covid-19 was mentioned on the death certificate, had Covid-19 as the underlying cause.  That’s lower than it’s been for quite a long time (apart from the previous week when it was barely higher, at 78%), but it was at roughly that level, or often quite a bit lower (going below 70% some weeks), between mid-March and the start of July last year, including periods when both Alpha and Delta were the dominant variant.  On the whole, the pattern is roughly that the percentage of Covid-related death registrations where Covid is the underlying cause is lower when overall Covid-related death registrations are relatively low, or when they are falling.

“Then, one thing to be clear on is that the distinction in the death registration data between deaths, where Covid is mentioned on the certificate, that do and do not have it classified as the underlying cause of death, is nothing like the well-known issue with the 28-day definition used on the Government dashboard.  With the 28-day definition, someone could die “with Covid” if they happened to be killed in a car crash within 28 days of a positive test, with their Covid status having nothing at all to do with their death, and it would be counted as a Covid death in the main dashboard figures.  With the registration definition, a doctor would not put Covid on the death certificate in those circumstances.

“The certificate has two sections about cause of death.  In one, the doctor lists the sequence of diseases and conditions that directly led to the death, and if Covid is in that part, it would usually be regarded as the underlying cause.  In the other section, the doctor lists “Other significant conditions contributing to death but not related to the disease or condition causing it.”  If they mention Covid there, it wouldn’t be coded as the underlying cause, but it would have to be ‘significant’ and ‘contributing to death’.  That could be the case if, for instance, if their Covid made a disease, that they already had, worse, so that it killed them much more quickly than would otherwise have happened.  So the person who gets infected, maybe has no symptoms, and dies in a car crash on day 27 after their positive test, would not have Covid mentioned on the certificate, because it didn’t contribute to death.  If Covid is mentioned on the certificate, but not as the underlying cause, that would mean that Covid hastened the death, or made the person’s last days much more difficult, or something like that.”




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



Declared interests

None received.

in this section

filter RoundUps by year

search by tag