select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
The Office for National Statistics (ONS) have released the latest figures for deaths in England and Wales, including deaths from COVID-19 in all settings.
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“There are good and bad aspects of the news from this week’s ONS release of data from death registrations. I’ll consider the more positive aspects first. Though registrations of deaths from all causes in the most recent available week (ending 25 September) are above the average level of the last five years, the number of excess deaths is not large, and the total number of deaths is less than the highest level for this week seen during those five years. So we’re not seeing a major surge in excess deaths. Also, the number of deaths for which Covid-19 is mentioned on the death certificate is still not high (at 215 registered in the latest available week). That’s the highest level since late July, but, as the ONS bulletin shows, it’s a lot smaller than the number of deaths in that week where influenza and/or pneumonia are mentioned on the death certificate (1,172 in the latest available week of registrations, and the Covid-19 deaths are not double counted in this total).
“However, the number of deaths with Covid-19 on the certificate has continued the rise that we’ve seen since early September. The data for September are consistent with the number of deaths doubling roughly every two weeks, though there’s a lot of statistical uncertainty about the doubling time. It’s hardly surprising that deaths have risen, given the other data that we’ve seen recently on increases in the numbers of infections (from the ONS Infection Survey and the REACT-1 study, for instance). Covid-19 can be a very serious illness in some people, and sadly a small proportion of infected people will die. The death registration data don’t yet show any of the signs of a slowing up in the rate of increase, as has been seen (as a distinct possibility) in data from the ONS Infection Survey and the most recent data from the REACT-1 study – but that’s also not really surprising. A certain amount of time, maybe two or three weeks or even more in some cases, has to pass between a person first being infected and later becoming ill enough to die. So the numbers of deaths towards the end of September would be most closely related to the numbers of new infections in early September, well before the signs of slowing in the rate of new infections that appeared later in the month. If these patterns continue to hold, we’d unfortunately expect to see a continuing increase in deaths from Covid-19 in the registration figures for the next few weeks at least, though perhaps with a slowing of the rate of increase during that time – but nobody can be sure of this yet.
“It’s worth mentioning that these numbers of death registrations cannot have been affected at all by the recent issues in recording new infections through the Test and Trace system. They are collected through an entirely different system, that is very robust and has been developed over many years.”
Prof Sheila Bird, Formerly Programme Leader, MRC Biostatistics Unit, University of Cambridge, said:
“Today’s provisional account of COVID-19 deaths by date of occurrence now shows two COVID-mention deaths in England prior to March 2020: on 30 January 2020, death of 84 year old male in the South East, as first reported by the Sun on 8 September (cause of death at coroner-autopsy having been revised in August 2020) and on 2 February 2020, death of a male aged 55-59 years in the Eastern region. I assume that this death had been referred to the coroner but whether review of autopsy findings or inquest-verdict accounts for the late registration remains to be explained. But do please explain.
“Equally notable is that currently COVID-19 deaths, by registration-week or earlier by occurrence-week, evidence a doubling time of 2 weeks. This is concerning, as was forecast.”
Prof Sir David Spiegelhalter, Chair of the Winton Centre for Risk and Evidence Communication, University of Cambridge, said:
“The latest data confirms that Covid deaths in the UK were doubling around every two weeks in September, although are still only comprised 2% of deaths in the country.
“Crucially, there is still absolutely no sign of any reduction in the 30% increase in deaths occurring at home – if this is to be a long-term feature of deaths in this country, we need to be confident that appropriate end-of-life care is being made available.”
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.”
None others received.