Some very dodgy maths in the @Mailonline suggesting that two thirds of those admitted with COVID recently were actually admitted for another reason.
It's nonsense, as all they have done is difference the figures between the two dates.
Let's take a look.
1/
It's nonsense, as all they have done is difference the figures between the two dates.
Let's take a look.
1/
From the primary diagnosis data from the NHS, I sort of agree the figures of 563 and 197, (except they've used the wrong starting date, but that's not the issue here).
Thus the Mail concludes that 366 (65%) were admitted for non-COVID reasons over 15 days.
Why is it wrong?
2/
Thus the Mail concludes that 366 (65%) were admitted for non-COVID reasons over 15 days.
Why is it wrong?
2/
It's wrong because it assumes no discharges over the period. Let's take a simple example to illustrate that.
In the scenario shown, the Mail would conclude that nobody has been admitted with COVID as the primary cause, and thus assume ALL new admis'ns were for other reasons.
3/
In the scenario shown, the Mail would conclude that nobody has been admitted with COVID as the primary cause, and thus assume ALL new admis'ns were for other reasons.
3/
In reality, there's a high turnover of patients across the period. So you might actually have this situation, which gets you to the same end point, but in which the actual % of new admissions with primary cause COVID is 73%, not 0%.
4/
4/
So the conclusion is that differencing the start and end dates gives no indication of the proportion of new admissions who have COVID as their primary cause.
@Mailonline - it's wrong -please remove it.
HT @jim_reed for alerting me to this.
@Mailonline - it's wrong -please remove it.
HT @jim_reed for alerting me to this.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
