...
If you are in hospital 29 days after testing positive for Covid then there seems a pretty good chance your death was Covid related. ...
They are recording all deaths that have COVID19 on the death certificate.
They're also recording the excess mortality rate.
The deaths in hospital after 28 days is pretty much just a statistical tool to get consistency.
When deaths in hospital have been used as the official rate for previous epidemics nobody has batted an eyelid; it's only because it's now affecting us that anybody cares.
This might lead to an international discussion about standardised measurements, but for planning I would have thought as much data as possible and consistency with presenting it are probably the most important things.
In the insurance world I don't think the PRA would give us a bye from producing a more accurate number just because it was inconsistent with some smaller number from some previous year. It just seems an unprofessional thing to do too. Just report it as accurately as you can.
You have two known errors. A large error that meant you understated past deaths and a small error that meant you overstated some past deaths. If you are correcting your number for past deaths correcting the first seems more important and is something that has more urgency. It wouldn't really bother me if they hadn't changed it, it is the big song and dance they made of discovering the second error; and then the urgent need to correct it after been happy to ignore the bigger error that would have made them look worse. That is what grates.
If you are correcting future errors then sure, you can make the change to future reporting. But even then you should be aware of the potential understatement going forward.
I'd hope the people doing the future planning are aware of the weekness in their modelling, then maybe it doesn't matter so much. If the right people are aware the real number is about 60k and not about 40k then guess we will be fine.
