Read on Twitter
The headline figure from today's new @ONS all-cause deaths data is (mildly) good news. Deaths were slightly lower in the week ending 27th November than the previous 7 days.
BUT
The decline is slight, and there are reasons we should still be worried...
BUT
The decline is slight, and there are reasons we should still be worried...
One big concern is that deaths from COVID-19 have actually continued to rise, they have just been offset by a fall in deaths from other causes.
Another concern is that deaths in the oldest, most vulnerable age group are still rising. The fall in this latest week of data has been driven by a drop in deaths among 75-84 year-olds.
This is not too surprising, given the patterns we've seen in the age-specific case data. Cases have fallen much more slowly in the older age groups. https://twitter.com/VictimOfMaths/status/1335970875997380614
There are also, as ever, quite stark regional differences in the trends. The North West seems to have turned the corner, as do the West Midlands, London, South East, Scotland and Northern Ireland.
But deaths in Yorkshire, the East Midlands and the South West have risen.
But deaths in Yorkshire, the East Midlands and the South West have risen.
One piece of slightly better news is that there hasn't been a rise in care home deaths, which is what you might immediately fear with higher case and mortality rates in the oldest age groups.
In fact, more up-to-date figures from the Care Quality Commission show that COVID-19 deaths in care homes are falling, albeit fairly slowly.
It's also interesting to contrast the other cause section of this graph between the Spring and Autumn 'waves'.
It's also interesting to contrast the other cause section of this graph between the Spring and Autumn 'waves'.
If we dig down further, there isn't any evidence that this data is missing large numbers of deaths of care home residents who are dying in hospital.
Again, interesting contrasts with the Spring.
Again, interesting contrasts with the Spring.
R code for the plots in this thread is here (for the CQC figures):
https://github.com/VictimOfMaths/COVID-19/blob/master/All%20Cause%20Mortality/COVIDCareHomeDeaths.R
and here (for everything else): https://github.com/VictimOfMaths/COVID-19/blob/master/All%20Cause%20Mortality/AllCauseDeaths.R
https://github.com/VictimOfMaths/COVID-19/blob/master/All%20Cause%20Mortality/COVIDCareHomeDeaths.R
and here (for everything else): https://github.com/VictimOfMaths/COVID-19/blob/master/All%20Cause%20Mortality/AllCauseDeaths.R
h/t to @danielhowdon for pointing the CQC data to me, which I'd previous missed.
It lives here if you want to have a look yourselves:
https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/datasets/numberofdeathsincarehomesnotifiedtothecarequalitycommissionengland/2020
It lives here if you want to have a look yourselves:
https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/datasets/numberofdeathsincarehomesnotifiedtothecarequalitycommissionengland/2020
