How many people in the UK have had Covid since Omicron came along?

[andrewducker]

I was curious as to how many people across the UK have had Covid since Omicron came along. So I went to the ONS Infection survey, which is based on them sampling the same people regularly, and checked what percentage of people each week have had Covid, across each nation of the UK since Omicron came along (I picked 20th of December as a cut-off date).

	Scotland	England	Wales	NI
20 December 2021	2.57	3.71	2.52	2.59
28 December 2021	4.52	6.00	5.20	3.97
03 January 2022	5.65	6.85	5.56	5.41
12 January 2022	4.49	5.47	3.69	5.68
19 January 2022	3.11	4.82	3.27	5.26
26 January 2022	3.52	4.83	4.57	7.43
02 February 2022	4.01	5.18	3.99	7.93
09 February 2022	4.17	4.49	3.71	7.99
16 February 2022	4.57	3.84	3.23	7.23
Total	36.61	45.19	35.74	53.49

And it's interesting to note that over half of the people in Northern Ireland have had Covid since Omicron, compared to 45% for England, and only 36% for Scotland.

Which might explain somewhat why Scotland's figures are still going up, while England's are going down. Or then again, it might be something entirely different!

Comments

[doug]

Uh, I'm not sure you can just add those figures up like that. This is for nose-and-throat PCR, and people can easily test positive for more than a week, so the 4.57% in Scotland on 16 Feb 2022 probably includes a significant chunk of the 4.17% who already had it on 9 Feb 2022.

It is, I guess, something of an upper limit ballpark figure, though. Although their model is for how many would test positive, not how many had it, and the latter of course would be bigger. (My recollection is sensitivity for nose-and-throat PCR in the field is estimated at 70%, but that is a very old figure.) No idea whether those two errors cancel out! I'd guess the true figure of infections would be somewhere lower than adding up weekly ONS rates, but that's not based on actually doing any sums.

If we're being really picky it's not accounting for reinfections, but even with omicron there's not a lot of them within two months.

A different and arguably more defensible approach would be to look at the dashboard stats for new cases reported and add those. That would give you a smaller value, since a lot of cases aren't detected - the awesome thing about the ONS survey is that it avoids that problem. Even worse, the case ascertainment ratio (i.e. how many are detected out of how many there really are) varies - there were a lot more cases uncovered around Christmas/New Year than at other points, probably because loads of people were doing LFTs before meeting up. There are good estimates for ascertainment around - the simple estimate is to benchmark against ONS prevalence estimates - but I don't immediately have my hand on them.

[rhythmaning]

I believe that some of the issues that Doug raises are why many graphs use two-week moving averages to record covid infections.

I recall that reinfections were beginning to be recorded (unbelievable that they weren't previously!), but I can't remember which data series it applied to!

And I appreciate the erudition of your commenters.

[mellowtigger]

Stay negative and help others stay negative for as long as you can. The decade-long prospects for anyone infected at any severity level are troubling (as a world-class understatement). The outcomes look even worse as people are reinfected. I know I'm one of the seeming tin foil hat voices on this issue, and I'm okay with that assessment. Like the climate crisis, "What if we make all of these positive changes for what turns out to be no valid reason?" I hope I'm wrong, but the chances of me being wrong seem to diminish with each new study that comes out.

[trailer_spot]

It's been two years since this whole thing started. And it's hard to believe how little real-time, reliable data/observation related to the pandemic exists here in Germany. I don't think we take such samples in regular intervals nationwide to get an overview in addition to the testing regime. The time lag until we know how many people are in intensive care in hospitals is also still considerable.