select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
The Office of National Statistics (ONS) have released their latest statistics for deaths from COVID-19 in the UK, now including the week ending in 17th April.
Prof David Leon, Professor of Epidemiology, London School of Hygiene & Tropical Medicine, said:
“The figures published today by ONS show clearly that there are now substantial numbers of people dying in care homes where a doctor who certified the death believes that the deceased either died because of COVID-19 or had COVID-19 at the time of death. These have increased sharply since the previous week.
“What do these new data add to our understanding of overall trends in COVID-19 mortality? They provide reliable numbers of all COVID-19 deaths in England and Wales from the start of March to April 17 by day of death regardless of where the death occurred. This shows that the peak in all COVID-19 deaths, regardless of whether they occurred in hospital or elsewhere, was on April 8. Since then there has been a steady decline. Secondly the data reported suggest that the number of COVID-19 deaths in care homes up to April 17 has been high but relatively stable since April 12, running at an average of just over 300 per day in this most recent reporting period. Thirdly, it shows that the small number of COVID-19 deaths occurring in people’s own homes has declined from a peak of 70 on April 5 to under 25 per day since April 15.
“These data have limitations. Firstly it is still likely that there will have been some deaths involving COVID-19 that occur in care homes and other places outside hospital that are missed. Secondly, the way in which a COVID-19 death is defined by ONS is that COVID-19 is stated somewhere on the death certificate. This does not automatically mean that such a death was actually caused by COVID-19, although the majority are likely to have been related to the viral infection. Most importantly, however, these numbers do not count the indirect deaths that are occurring because people with serious symptoms or diagnoses of non-COVID health conditions are not getting diagnosed in a timely way or are experiencing delays in treatment. Some of this will be due to people’s reluctance to go to Accident and Emergency departments, while in other cases it is because investigations or treatments are being scaled back by hospitals.
“The new report from ONS gives some insight into the potential scale of these indirect deaths. It shows that the number of excess deaths (compared to what would be expected based on the past 5-years) in the week ending April 17 (11854) is considerably greater than the number of COVID-19 deaths in the same week (8758). Some of this difference may be because of undiagnosed COVID-19 deaths. We will get a much better idea of how far this is true when the ONS releases its next report on May 5. This will enable us to see whether in the week ending April 24 there has been a decline in the difference between COVID-19 deaths and total excess deaths. This would be expected if much of the difference is due to undiagnosed COVID-19 deaths. The reason for focussing on next week’s data is that it will be for a week over which there has been a clear downward trend in hospital deaths following the April 8 peak.
“So far the ONS data per se do not support concern that the numbers of COVID-19 deaths are spiralling out of control in care homes at least up until April 17. However, we need to see trends in numbers of all cause deaths in care homes and elsewhere in next week’s release to be able to see more clearly what the situation really is.”
Prof Sir David Spiegelhalter, Chair of the Winton Centre for Risk and Evidence Communication, University of Cambridge, said:
“The latest data from ONS show that last Saturday’s stories about passing 20,000 COVID deaths were somewhat late – just in England and Wales, the total went past that milestone on 15th April.
“The latest data from ONS report 22,351 deaths registered in the week ending April 17th, a huge excess of 11,854 over the average for the last five years. Of these extra deaths, 74% were labelled as COVID, leaving around 3096 (26%) extra ‘non-COVID’ deaths. This proportion is larger than last week, suggesting these are not just due to early reluctance to put COVID on a death certificate without testing.
“There were 11,886 registered deaths in care-homes and at home, compared with the average 4,498 for this time of year. Only 2,466 of these were COVID deaths, leaving nearly 5,000 extra non-COVID deaths occurring out of hospital. Some of these will be undiagnosed infections, and around 1,400 non-COVID deaths were ‘exported’ from hospitals to the community. But presumably the massive drop in hospital attendances has impact, and many of the 5,000 would have been alive had they received standard health-care. It seems vital to get the NHS back to business, but this will be difficult if people see hospitals as dangerous places.
“While COVID deaths in hospitals have been steadily declining since April 8th and now number around 400-450 a day, new Care Quality Commission data show that last week there were around 350-400 COVID deaths in English care-homes each day. When we add in deaths at home, this suggests there are now about as many COVID deaths out of hospital as in hospital. And while hospital deaths are steadily decreasing, there is no sign yet that we are past the peak in care homes.
“In the previous week the CQC data closely matches ONS death registrations, suggesting that their recent data will later be reflected in the official death registrations.”
Nigel Marriott FRSS, Independent Statistician, said:
“I expect there to be a greater focus on the excess mortality figures published every Tuesday by the ONS going forward which have been at record levels for the last 3 weeks. The Financial Times produced a very interesting article at the weekend comparing excess deaths across 14 countries with nearly countries all showing a similar picture to the UK in terms of excess deaths being considerably higher than reported COVID19 deaths.
“Unfortunately excess death figures are published 10-17 days in arrears. The time lag is unavoidable for data quality reasons so I am interested in seeing if it is possible to estimate what these figures will be up to 17 days ahead using the daily data published by NHS England & DHSC. For the week ending 17th April, the excess of 11395 for England was 2500 higher than my estimate and 1000 higher than the estimate produced by Chris Giles of the Financial Times. That indicates higher than expected deaths outside of those due to or with COVID19 in hospital which is concerning and will require further in-depth analysis. We need to know if this difference is due to underreported COVID19 deaths or are in fact deaths due to other causes and the answer will have implications for decision makers.
“Was this the peak week for reported excess deaths? Currently I am estimating reported excess deaths for England up to 24th April to be around 9000, 2400 lower than the figure reported today, so I think this week’s figures are likely to have been the peak. If next week’s figures end up being higher than this week’s, that will be extremely concerning and would make it more difficult for decision makers over when and how to relax the current lockdown.”
Prof Sheila Bird, Formerly Programme Leader, MRC Biostatistics Unit, University of Cambridge, said:
“Due to registration-delays beyond 8 days (unlike Scotland), COVID-mention deaths that actually occurred in week 16 and are reported today [7 288] are a substantial under-count for the number that will eventually be registered as having occurred in week 16.
“71% of the 26 816 additional deaths (that is: above expectation based on past 5-years) that have been registered in E&W in the COVID-era (registration weeks 11-16) have been COVID-mention deaths [18 446].
“Deaths in E&W that occurred during weeks 11 to 15 included only one COVID-mention death-under 15 years of age among 1 221 COVID-mention deaths aged under 60 years. Another already-registered COVID-mention death under 15 years of age has occurred in Week 16, one of 591 already-registered COVID-mention-deaths under 60 years of age that occurred in week 16 in E&W. Some COVID-mention-deaths under 15 years of age, because so unusual, may be referred for inquest and hence would not ordinarily be registered with the Office for National Statistics until the inquest-verdict had affirmed cause of death.
“The additional material provided by ONS highlights the important of tracking care-homes’ COVID-mention deaths by death-date because reporting delays are substantial to Care Quality Commission and to ONS but ONS’s definition catches more COVID-related deaths in care homes. Worryingly, the trajectory of these deaths has not plateaued on the basis of those that occurred by 17 April and were registered by 25th April 2020.”
Declared interests
Prof Sheila Bird: SMB leads for the Royal Statistical Society on the need for legislation to end the late registration of deaths in the UK outside of Scotland.
None others received
https://content.govdelivery.com/accounts/UKONS/bulletins/288c4d7
All our previous output on this subject can be seen at this weblink: www.sciencemediacentre.org/covid-19/