expert reaction to ONS stats on deaths registered weekly in England and Wales, provisional: week ending 21 May 2021

expert reaction to ONS stats on deaths registered weekly in England and Wales, provisional: week ending 21 May 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 21 May 2021.


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

“This week’s release from ONS of death registrations takes the data up to the week of 15-21 May, so we’re eventually out of the disruption to the figures caused by registration delays at the time of the May Day bank holiday. Those delays made the number of registrations two weeks earlier, the week including the holiday, too low to represent when people actually died, and they also had a smaller effect on the registrations one week earlier, which would have been a little too high because of registrations moved from the bank holiday week. So we’ve got this week’s data clear of bank holiday effects, and next week’s will be too, but the week after that the data will include the late May bank holiday and will be disrupted again. ONS also provide data analysed by the date on which the deaths actually occurred, not when they were registered, but those figures are always incomplete for the most recent week (and, to a lesser extent, for a few weeks before that), so they can’t tell us much about short-term recent trends.

“On deaths involving Covid-19, the picture is very positive. In the most recent week, 107 deaths were registered in England and Wales that had Covid-19 on the death certificate, and 66 had Covid-19 classified as the underlying cause of death. The corresponding figures for the week before were 151 and 108. The weekly number of registered deaths with Covid-19 as the underlying cause hasn’t been as low as 66 since the week ending 13 March 2020, which was the first week that deaths involving Covid-19 were registered in England and Wales, with just 5 registrations. (There was one more week at the start of the pandemic, and another three in late August and early September last year, when the number of registered deaths with Covid-19 on the certificate, whether as underlying cause or not, was slightly below the latest week’s figure of 107.) When you think back only as far as the start of April when there were over 300 registered deaths in a week with Covid-19 as the underlying cause, let alone to late January when there were over 7,500 in a week, this is huge progress. Vaccination will have played a major part in this reduction, though lockdowns and other interventions played an important part too. People who, sadly, died in the latest week of those data would probably have been infected three or more weeks before, so well before the latest loosening of lockdown in England on 17 May – so it wouldn’t be possible to see any effect of that removal of restrictions on these numbers.

“The picture on total registered deaths from all causes is also positive, though a bit less encouraging. All-cause registered deaths in England and Wales were 3% down in the latest week compared to the week before, and were 328 below the average for 2015-2019. That’s now a run since the week ending 11 March where total deaths have been below the five-year average, apart from the week before the latest one when there were still some effects of the May Day bank holiday – so again no excess deaths this week. However, they aren’t all that far below the five-year average compared to the position in April. In Wales and in all five of the English regions in the North and Midlands, the number of all-cause deaths registered in the most recent week, ending 21 May, is higher than it was three weeks before that, in the week ending 30 April, just before the bank holiday disruption to the figures. Those increases certainly aren’t directly due to deaths involving Covid-19, which were down in every English region over that three-week period, and unchanged in Wales (though the numbers are tiny in Wales – just 7 deaths with Covid-19 on the certificate registered in the latest week as well as three weeks before). Also, over that three-week period, deaths involving Covid-19 fell in almost every age group. So the trends in all-cause deaths must involve other causes of death, and we don’t have information on the other causes of death yet. It might be relevant that the increases in all-cause deaths over that three-week period seem to have been mostly in younger age groups, up to people in their 50s. I can speculate that this might have something to do with earlier removals of lockdown restrictions, so that deaths from non-Covid causes that were reduced by lockdowns (such as deaths in road accidents) might have gone up a bit – but I really can’t say without more data on the causes. The pattern isn’t yet one of any great concern anyway.

“Returning to deaths directly involving Covid-19, there has been some concern that they might increase, because the numbers of new confirmed cases and of admissions to hospital on the dashboard at have recently been showing increases. Earlier in the pandemic, any sustained increase in new cases or hospital admissions would lead inevitably to increases in deaths a few weeks later. That process won’t occur in the same way now, because of the effect of vaccination on reducing serious illness, though it’s not yet entirely clear what the new pattern might be, or how new variants might play a role. But in any case, this new release of death registration figures can’t throw any light on that. The increases of new cases and hospitalisations on the dashboard didn’t really occur until after mid-May, and that wouldn’t have led to any noticeable increase in death registrations by the latest week in the data, 15-21 May.”



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



Declared interests

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  I am also a member of the Public Data Advisory Group, which provides expert advice to the Cabinet Office on aspects of public understanding of data during the pandemic.  My quote above is in my capacity as an independent professional statistician.”