The Guardian had a headline earlier:
Most UK deaths with Covid recorded in day since March
The number of people who died in the UK within 28 days of a positive Covid test, whose death was recorded in the last 24 hours, is at its highest since March.
Official data shows 146 people have been recorded as dying within the past day – the highest since 175 were recorded on 12 March
But it went on to say:
For the week ending 23 July, of 9,100 total deaths over seven days there were 308 fatalities in which Covid was mentioned on the death certificate.
Nobody knows what percentage of the population has had Covid, but in February they were saying that 15% of the population had contracted it - https://news.sky.com/story/covid-19-15- ... y-12207328
That's just less than 1 in 6 of the population, and by definition the number must have increased since January. It seems reasonable to assume that it's now at least 20%, or 1 in 5.
So if there were 9,100 deaths in a week, one would have expected at least around 1,800 of them to have tested positive for Covid. In fact, as its concentration is highest in the over 80's, and they are the large majority of the deaths one would have expected the figure to be well over 2,000.
Yet it was only 308. And in that context the latest figure of 146 seems neither here nor there. Even assuming that figure was replicated every day the weekly total would still be well under the 1 in 5 that one would expect from a random selection of people.
It therefore seems to me that the numbers of people dying with Covid, let alone of it, are now so small as to be not worth reporting, let alone reporting with a splashy headline (though admittedly, it was the Guardian!)
In fact, on these figures, it would seem that you're less likely to die if you have got Covid than if you haven't!
Got a credit card? use our Credit Card & Finance Calculators
Thanks to jfgw,Rhyd6,eyeball08,Wondergirly,bofh, for Donating to support the site
"Deaths with Covid"
Forum rules
This is the home for all non-political Coronavirus (Covid-19) discussions on The Lemon Fool
This is the home for all non-political Coronavirus (Covid-19) discussions on The Lemon Fool
-
- Lemon Quarter
- Posts: 2874
- Joined: November 6th, 2016, 9:58 pm
- Has thanked: 1389 times
- Been thanked: 3805 times
-
- The full Lemon
- Posts: 16629
- Joined: October 10th, 2017, 11:33 am
- Has thanked: 4343 times
- Been thanked: 7536 times
Re: "Deaths with Covid"
I think I agree with you. It is time (as I said a while ago) for these numbers not to be reported any longer as they are pretty meaningless.
Dod
Dod
-
- Lemon Half
- Posts: 8411
- Joined: November 8th, 2016, 5:56 am
- Has thanked: 4486 times
- Been thanked: 3619 times
Re: "Deaths with Covid"
Can I agree that "newspapers" need to find something else to talk about.
Reporting the data through the official channels/dashboard is probably still warranted; though the "deaths with 28 days of your first Covid test" is probably well past its sell by data (re-infection rates making it more moot as time goes on)
But....
It's important to pay attention to tenses though lest we fall foul of base rate fallacies
- the 15% of the population mentioned the linked article were mentioned has having had Covid by mid-January (i.e. cumulatively up to that point)
- it's having COVID that's the risk not having had it previously
If you were looking to infer whether a proportion of deaths looked significant you would need to look at the instantaneous prevalence of COVID - which in England is estimated at 1 in 75 presently
So, using the figures provided, if there were 9100 deaths in a week you would expect 121 of them to be COVID deaths if having COVID and dying were completely independent
But yeah, they should find something else to publish; statistics refreshers or grammar perhaps?
- sd
Reporting the data through the official channels/dashboard is probably still warranted; though the "deaths with 28 days of your first Covid test" is probably well past its sell by data (re-infection rates making it more moot as time goes on)
But....
It's important to pay attention to tenses though lest we fall foul of base rate fallacies
- the 15% of the population mentioned the linked article were mentioned has having had Covid by mid-January (i.e. cumulatively up to that point)
- it's having COVID that's the risk not having had it previously
If you were looking to infer whether a proportion of deaths looked significant you would need to look at the instantaneous prevalence of COVID - which in England is estimated at 1 in 75 presently
So, using the figures provided, if there were 9100 deaths in a week you would expect 121 of them to be COVID deaths if having COVID and dying were completely independent
But yeah, they should find something else to publish; statistics refreshers or grammar perhaps?
- sd
-
- Lemon Half
- Posts: 8411
- Joined: November 8th, 2016, 5:56 am
- Has thanked: 4486 times
- Been thanked: 3619 times
-
- Lemon Slice
- Posts: 785
- Joined: April 9th, 2021, 5:54 pm
- Has thanked: 225 times
- Been thanked: 265 times
Re: "Deaths with Covid"
Clitheroekid wrote:So if there were 9,100 deaths in a week, one would have expected at least around 1,800 of them to have tested positive for Covid.
In the last 28 days?
Weekly positive test rate is around ~200k, so ~800k in the last 28 days.
That is around 1.2% of the population.
Return to “Coronavirus Discussions”
Who is online
Users browsing this forum: No registered users and 26 guests