select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
The Office for National Statistics (ONS) have released their latest data for deaths in England Wales, including COVID-19 deaths, up to the week ending 8th May 2020.
Prof Sir David Spiegelhalter, Chair, Winton Centre for Risk and Evidence Communication, University of Cambridge, said:
“The new ONS data for England and Wales includes Friday Bank Holiday 8th when there were almost no registrations, so we must take care in not over-interpret trends.
“There are apparently no non-COVID excess deaths, which is in stark contrast to previous weeks of the epidemic and may be attributable to fewer registrations.
“There has been a continued shift away from hospitals: half of the registered COVID deaths were in the community. 39% of all deaths in care-homes were registered as COVID.
“The median age of registered COVID deaths rose to 87 for women, 82 for men, reflecting the death-rate in care-homes.
“1 in 1,600 of population have died of COVID over 6 weeks, but this disguises rates that vary more than 10,000-fold over different age-groups.
“1 in 70 of all over 90s (more than 7500) have now died with COVID.
“Only 1 of 7 million school-children aged between 5 and 14 has had COVID on their death certificate, although this will probably increase as some further cases may be under coroners’ investigations.
“Hospital disruption is reflected by there being only around half the normal number of non-COVID deaths in hospital, ‘exporting’ these deaths into the community.
“One group that has fared better during the epidemic is young men aged between 20 and 24. We would normally expect around 106 deaths to be registered over the peak 6 weeks of the epidemic, but only 69 have occurred, a 35% reduction.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“There are many encouraging aspects of this week’s data from ONS on death registrations. The number of deaths from all causes has continued to fall, as has the number where COVID-19 was mentioned on the death certificate – and this is true not only for the whole of England and Wales, but also for Wales on its own and for all the English regions considered separately. It also applies to numbers of deaths in different types of location (hospital, home, care homes, etc.).
“As with this data release in previous weeks, there are always questions to ask about decisions made by doctors in recording causes of death on the death certificate, and about deaths that might have occurred because of the lockdown and changes in the availability of healthcare, rather than from someone being actually infected with the virus. That’s why many people, including me, would routinely recommend that we look at excess deaths – that is, at how deaths this year compare with deaths last year, or with average numbers of deaths over a few previous years. ONS have done their usual thorough job and provided this information too. Excess deaths in the last week available here, ending on 8 May, were about 3,100, which is far less than the peak of over 12,000 a week in mid-April. One interesting point, though, is that the number of deaths where COVID-19 was mentioned on the death certificate, about 3,900 in the week ending 8 May, is larger than the total of excess deaths for that week. That is, the number of deaths where COVID-19 was not mentioned was 800 fewer than the average number of deaths that week for the previous 5 years. I can’t say why this is, and it will take considerable further analysis from ONS in the future to throw light on it.
“Two possible explanations did occur to me, though I can’t say how likely they are. One is that the week ending 8 May included the early May bank holiday. Bank holidays, according to ONS, always change exact dates of death registration. It’s difficult to assess the impact of that this year – the holiday was on a different day of the week than usual, and the impact on registration might well have been affected by the ongoing lockdown and also by the fact that it was a ‘special’ bank holiday to celebrate the anniversary of VE Day. What’s more, registry offices were allowed to take registrations over the holiday weekend this year because of the special circumstances of the pandemic; that usually doesn’t happen (though in fact only 88 deaths were registered on the actual bank holiday). ONS say that the effect of the bank holiday may make trends less clear for the data release of this week and next week, and of course there’s another bank holiday soon, but we’ll have a clearer picture before long. Another possible explanation is more subtle. Perhaps all the public discussion about issues of mentioning COVID-19 on death certificates, and the importance of looking at excess deaths, might have changed what doctors record on certificates where it’s not completely clear whether COVID-19 is involved. The rules on what should be entered on the certificate look clear, but there is always a certain amount of professional judgement involved. Maybe the way that judgement works has changed, consciously or unconsciously, to make it more likely that COVID-19 is mentioned than it would have been a few weeks ago. I have no evidence that this has actually occurred – that will need more detailed investigation, and we may well never find out. But if it has occurred to some extent, maybe that’s a partial explanation of why there were fewer non-COVID-19 deaths in the week ending 8 May than the previous 5-year average.
“The position on care home deaths in this release hasn’t moved on a great deal from the much more detailed report on deaths to care home residents, issued by ONS last Friday1. That was based on deaths registered up to 9 May, so one day more than today’s release, though the analysis is mostly only for deaths that actually occurred up to 1 May, to take account of possible registration delays. That report showed that deaths of care home residents where COVID-19 was mentioned, and also excess deaths of care home residents where it was not mentioned, had both fallen very considerably since their peak in mid-April, and today’s release gives much the same message. It is still concerning that such a large proportion of deaths in care homes continued to involve COVID-19 – 40% of all deaths in care homes in the week ending 8 May. However, the actual number of such deaths in that week, 1,666 in England and Wales, is well down on the peak weekly number of nearly 2,800 in the week ending 24 April. Even if the figure for the week ending 8 May is eventually revised upwards a bit because of late registrations, as seems quite probable, the numbers are moving in the right direction.”
1. Deaths involving COVID-19 in the care sector, England and Wales: deaths occurring up to 1 May 2020 and registered up to 9 May 2020 (provisional). https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/articles/deathsinvolvingcovid19inthecaresectorenglandandwales/deathsoccurringupto1may2020andregisteredupto9may2020provisional
All our previous output on this subject can be seen at this weblink:
www.sciencemediacentre.org/tag/covid-19
Declared interests
Prof Kevin McConway: “Prof McConway is a member of the SMC Advisory Committee, but his quote above is in his capacity as a professional statistician.”
None others received.