Thread by @gnomeicide

Covid data has gone bonkers. I wasn't planning to do a thread today, but what the hell, why not. Here's the overall picture before some more detailed analysis. (1)

Deaths continue to fall ever so slowly, which is a remarkable testament to the incredible inroads made with vaccines and treatments - by now cases should be rising, following cases haven risen for weeks (2)

But 111 deaths a day remains terrible. And the rate of fall, under 1% a day, is barely perceptible (3)

The reason they're falling is science. New treatments and of vaccines, especially boosters right now. The proportion of recorded infections translating into deaths is falling. 0.247% - the lowest since Delta took over, and the fall coincides precisely with boosters (4)

You'll notice if you just put the daily deaths data on a graph that there are two trends - week days (Tues-Sat) and weekends (Sun-Mon). So yesterday and todays data don't mean so much. Tuesday is the day when your eyes pop out at the deaths data (5)

Here's the same data showing all of those previous points more clearly - weekend data matters less, the fall is slow but fairly steady. (6)

Cases are bonkers. 78229 daily average over the last 7 days, no obvious sign of that rise abating yet (7)

Remember, that DOES NOT include reinfections. We're up over 11 million recorded cases in the UK now, so thats a LOT of reinfections not getting reported. And remember the chronic shortage of test strips too - we're seeing this rise DESPITE those bottlenecks (8)

Looks even more bonkers in raw form. (9)

Last Sunday the 7 day rolling case total was 360,480. After yesterday it was 547,693. So thats a rise of 187,123. A 52% increase in a week. 6% a day. Doesn't take a genius to say that deaths continuing to fall with this going on is a big, big ask (10)

Admissions data lags, we only have that until the 13th of December, and admissions are rising (11)

With cases rising so very fast its hard to see how, even with new treatments coming online today, that doesn't continue. Where case data goes, hospitalisations follow (12)

And most often when hospitalisations go up, deaths follow (13)

That deaths are not yet rising yet shows the astonishing might of vaccines and improved treatment. But from where we are a small daily percentage change in hospitalisations lands us in trouble, and FAST (14)

There is no precedent for a massive rise in infection that has not led to an NHS crisis. Can we do better this time? As a proportion hospitalisations are lower than ever (15)

And maybe that will fall again. But the rate of cases rising is stratospheric. To continue as we are without further restrictions is one hell of a roll of the dice. (16)

I want to tell you to go and enjoy your Christmas. If you give your Mum Covid on Christmas
Eve, on average she'll show symptoms on the 29th. If she's hospitalised, on average thats around New Years Day (17)

And if it kills her the peak day for deaths would be the 16th of January. Last years peak for deaths shows that the period from Christmas to just before New Year was when most people caught Covid then (18)

...folk who went home for Christmas just before the restrictions hit, they gave Covid to their old folks over Christmas and mostly went back to where they live to spend New Year with younger mates or alone (19)

...and by the middle of January hospitals were swamped, and peak deaths came exactly when we expected. We actually don't know when peak hospitalisation and death from Omicron will occur, but its unlikely to be very different (20)

...and short of a medical miracle, its hard to see how we avoid that now. Even if new restrictions are announced we are going to see that rush to travel home. We -are- in trouble. Political dithering has, once again, turned a problem into a massive human tragedy (21)

Take care of yourselves and each other, folks. Our government has abandoned you. (fin)

Addendum: We can work out R (infection rate) from trends in deaths, hospitalisations and cases. And the three parameters pretty much track, except when new treatments etc. arrive. (a)

But of course the delays at which we can calculate the three are different - cases gives us data a week ago (13th Dec), hospitalisations further back (3rd Dec) and deaths further back again (26th Nov). Look carefully at the recent trends (b)

Anecdotally we're hearing of chaos in hospitals, with admissions already rising. We'll see whether the data tracks by the end of this week. We have a little longer to see whether deaths track with the recent stellar rise (c)

But infection is spreading phenomenally fast, and the danger here, if not apparent from the other data presented, MUST show here. (fin)