Issues with Mortality Analyses when the Data is Incomplete.
The official UK statistics agencies don't think there is a problem.
Many of us are trying really hard to report the COVID facts using the most reliable data we can access. Unfortunately, much of this data is not readily available. We have to use FOI requests and bespoke data requests (that we pay for out of our own pockets), simply to get useful data to work with.
The data that the ONS releases and indeed uses for their mortality reports is flawed. I have explained this to them and to their regulator, the UK Statistics Authority.
They do not seem to think that there is a problem.
I beg to differ. As usual, I resort to empirical data to demonstrate (how ironic).
These are the results of the only credible mortality data series ever created by the ONS. It is the result of a bespoke request I made several months ago and was updated a couple of weeks ago.
It includes all deaths that occurred up to 31-Jul-22 that were registered by 11-Nov-22 or thereabouts (it’s not entirely clear).
If I group the deaths by 5-year age ranges for convenience, for all ranges above 75 years old, you can see using your logical eye that the series look reasonably complete:
However, for all the younger ages, you can see a progressively more significant tailing off in deaths in 2022. And that’s just up to end of July which is already almost four months after the report was produced!!
The question is, are the younger ages suddenly dying less or is there really an issue with registration delays??
How does it look to you? How reliable do you think any analysis of mortality in England might be that includes 2022 data? I estimate that there are thousands of unreported deaths in this data series, and being loaded to the younger ages, in relative terms this could be quite important if there is a fatal signal somewhere in the data.
If you were in charge of the national statistics authority or its regulator, would you be “satisfied” with any of this? Would you care?