High positivity rate - why?

Get your science fix here: research, quackery, activism and all the rest
Post Reply
User avatar
Brightonian
Dorkwood
Posts: 1429
Joined: Mon Nov 11, 2019 3:16 pm
Location: Usually UK, often France and Ireland

High positivity rate - why?

Post by Brightonian » Mon Dec 27, 2021 5:19 pm

49.7% positivity rate in Ireland. What might be behind that greatly increased percentage? Perhaps a really huge proportion of the population has been infected with Omicron, and perhaps normally tests are performed on the symptomless for routine reasons (travel etc.), and they're found to be positive?

monkey
After Pie
Posts: 1906
Joined: Wed Nov 13, 2019 5:10 pm

Re: High positivity rate - why?

Post by monkey » Mon Dec 27, 2021 5:39 pm

Brightonian wrote:
Mon Dec 27, 2021 5:19 pm
49.7% positivity rate in Ireland. What might be behind that greatly increased percentage? Perhaps a really huge proportion of the population has been infected with Omicron, and perhaps normally tests are performed on the symptomless for routine reasons (travel etc.), and they're found to be positive?
People might not be doing the routine tests for work and whatnot over Christmas too, because they're not working and whatnot.

Someone may have just made a mistake too. Or something silly, like results weren't reported for a few days, and we're logged on that day, rather than the day of the test.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Mon Dec 27, 2021 5:42 pm

All the data is apparently linked to here.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
Brightonian
Dorkwood
Posts: 1429
Joined: Mon Nov 11, 2019 3:16 pm
Location: Usually UK, often France and Ireland

Re: High positivity rate - why?

Post by Brightonian » Tue Dec 28, 2021 8:16 am

I've found elsewhere a suggestion that only people who really, really wanted to get tested (i.e. with a bunch of symptoms), got tested on Boxing Day instead of using up that day for routine testing. Makes sense.

temptar
Fuzzable
Posts: 282
Joined: Sun May 10, 2020 6:19 pm

Re: High positivity rate - why?

Post by temptar » Tue Dec 28, 2021 9:16 am

You cannot book a PCR test for love nor money in Ireland at the moment. Slots just aren’t available. I suspect most people looking for one have had positive antigen tests because people have been doing them in advance of family events too. So the selection criteria are somewhat weighted.

I guess a lot of potential negatives were already weeded out. Omicron is around 90% of cases here at the moment too and anecdotally, I know a lot of people I cannot see this trip because they are isolating because of positive cases close to them.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Tue Dec 28, 2021 11:30 am

shpalman wrote:
Mon Dec 27, 2021 5:42 pm
All the data is apparently linked to here.
Where on there or here can I either find the numbers behind that 49.7% figure, or the actual number of tests per day? (It says 13149 were done in the last 24 hours but not the whole time series. The graphs are the total tests done, not the tests done each day.)

... here it does currently say that the latest daily cases, from yesterday (the 27th) are 6735 but only that 231,606 tests were done in the last 7 days for a positivity rate of 27%.

Zooming in on the cases per day graph (what moron plots deaths per day on the same scale?) we can see that they spiked a couple of days earlier, at nearly 14,000.
Screenshot 2021-12-28 122835.png
Screenshot 2021-12-28 122835.png (41.91 KiB) Viewed 2211 times
So I'd guess that what looks like a drop in cases is actually a drop in testing over the past couple of days.

Of course if the tests per day data were more easily visible we could see if the spike in cases was actually caused by a spike in testing, from lots of people wanting to meet up and getting lateral flows for example.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Tue Dec 28, 2021 11:42 am

COVID-19 Laboratory Testing Time Series.csv seems to only get as far as the 22nd of December.

So there's nothing to explain, I consider the report of 49.7% false until I can find the data which says it.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

temptar
Fuzzable
Posts: 282
Joined: Sun May 10, 2020 6:19 pm

Re: High positivity rate - why?

Post by temptar » Tue Dec 28, 2021 12:14 pm

https://mobile.twitter.com/COVID19DataIE

This picks up figures daily from the link you found above and the numbers you are looking for are there. Those figures were official figures.

Note: there is a lack of testing capacity compared to demand at the moment. There is probably a tighter selection of candidates for tests. Antigen tests may have weeded a lot of candidates out before they get to the central testing system. In the last week or ten days there have been issues with a backlog too.

Today’s figure is around 30% for the record. Ireland starts vaccinating children this week apparently.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Tue Dec 28, 2021 12:43 pm

So if I want to make my own spreadsheet I have to type the numbers out of the images on twitter?
covid.png
covid.png (18.26 KiB) Viewed 2192 times
But yes if we'd found this in the first place (silly me, expecting such a thing to be somewhere on the official government portal) it would have been obvious that the high positivity rate corresponds to a lot fewer tests compared to e.g. the 24th and the number of positive tests reported on the 27th is anomalously low. Even today it's probably still quite a bit lower than it should be.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

temptar
Fuzzable
Posts: 282
Joined: Sun May 10, 2020 6:19 pm

Re: High positivity rate - why?

Post by temptar » Tue Dec 28, 2021 1:10 pm

Yesterday and today are bank holidays. There are countries which didn’t even release numbers across Christmas Day including the UK. I don’t have a computer to hand so I can’t help you navigate figures and how the stats are managed.

That being said, you are the one who did not want to believe Irish media sources on the question.

I gave you quick access to the numbers. They were not false as you asserted.

Those numbers were published as official figures on pretty every media channel.

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Tue Dec 28, 2021 2:36 pm

temptar wrote:
Tue Dec 28, 2021 1:10 pm
Yesterday and today are bank holidays. There are countries which didn’t even release numbers across Christmas Day including the UK. I don’t have a computer to hand so I can’t help you navigate figures and how the stats are managed.

That being said, you are the one who did not want to believe Irish media sources on the question.

I gave you quick access to the numbers. They were not false as you asserted.

Those numbers were published as official figures on pretty every media channel.
And I'm also the one who actually looked at the numbers and answered the question implicit in the OP, that the high positivity rate corresponds to fewer tests being done and not because of tests "performed on the symptomless for routine reasons (travel etc.), and they're found to be positive".

Thanks for posting the link to the twitter which presumably scrapes the numbers off the official portal each week day. Actually it turns out that if you download the csv of the Testing time series off the portal it's fully up to date, it's just the preview which only goes up to the 22nd. And it also includes the weekends, which the twitter doesn't. So here's a plot of that:
covid-2.png
covid-2.png (33.58 KiB) Viewed 2162 times
The testing numbers aren't actually the same except for today, but there's presumably some protocol for correcting/updating the numbers for older dates as data comes in later? (The "Test24" column in the csv always corresponds to the difference between "TotalLabs" today and yesterday even if that's not the number which was on twitter that day.)

There's only a "Pos7" column corresponding to the new positives in the past 7 days, not a Pos24 corresponding to the past 24 hours, but I can take the difference between two rows in the Positive column for myself. Also here, the older numbers aren't the same as those given on twitter.
The Department of Health was today notified of a further 6,735 cases of Covid-19 in Ireland. The positivity rate is 49.7% - meaning almost half of the swabs taken were positive.
Actually, yesterday's twitter said 6,539 cases, and the csv currently suggests 6,578 yesterday, so where is "6,735" from? Are some of the cases in that total updates to older days? Where would I find an official source for "6,735"?

It's also misleading to quote that number without the context of the 7-day average having been heading up to 9,000-10,000 per day (i.e. this case number is a lot lower than usual) and without pointing out that only 13,419 (twitter) or 13,617 (csv) tests were carried out compared to the more usual ~35,000-40,000 last week (i.e. a lot fewer tests than usual) both of which pieces of information would have helped answer the question in the OP immediately.

(Twitter cases / twitter tests = 49.73% positivity rate for that day, CSV cases / CSV tests = 48.31%.)

(If you look at the UK's official covid portal now you can see that even if they didn't update it over the Christmas weekend, case numbers for England have been filled in for each day already; testing data never updates over the weekend and always reports the previous day so that will catch up this evening.)
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Tue Dec 28, 2021 4:10 pm

It occurred to me I can check the preview csv as compared to the updated one I downloaded.

Preview:

Code: Select all

2021-12-22T11:00:00	3273144	9694628	6421484	698468	7.2	33367	226737	38197	16.8	645
2021-12-21T11:00:00	3267118	9661261	6394143	691474	7.2	34981	229644	36027	15.7	644
2021-12-20T11:00:00	3256991	9626280	6369289	685790	7.1	27032	221814	34052	15.4	643
2021-12-19T11:00:00	3250099	9599248	6349149	680999	7.1	29155	222165	33386	15	642
...
Downloaded just now:

Code: Select all

...
2021/12/22 11:00:00+00	3279936	9701419	6421483	698934	7.2	38503	233532	38666	16.6	645
2021/12/21 11:00:00+00	3268773	9662916	6394143	691580	7.2	35504	231303	36136	15.6	644
2021/12/20 11:00:00+00	3258123	9627412	6369289	685864	7.1	27031	222950	34129	15.3	643
2021/12/19 11:00:00+00	3251232	9600381	6349149	681073	7.1	29459	223302	33463	15	642
...
So yeah previous numbers get retrospectively updated.

The UK makes a clear distinction between cases by date reported (which almost never gets retrospectively corrected) and cases by specimen date (for which the most recent few days are incomplete). Recently I discovered that it was possible to download any dataset, even e.g. cases by specimen date or deaths by date of death as complled on different dates, so that you can reconstruct the evolution of the cases or deaths on a particular day as more and more data comes in.

Italy just reports a number every day (which almost never gets retrospectively corrected) which is usually taken to be the data from the previous day, but sometimes there are notes and corrections in the csv files. (I don't bother with the official Italian covid cases portal, just the vaccination one.)
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Sun Jan 09, 2022 2:45 pm

I wanted to update this to see how things have been going in Ireland, since I feel like I've been hearing that their positivity rate is now high even on doing lots of tests.

After a couple of minutes of googling and following links around I gave up and came back here to find the link to the csv of the data because I don't remember how I managed to find it the first time.

The preview is still stuck at 22nd December but the downloaded csv now seems to stop at 6th September 2021.
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

User avatar
jimbob
Light of Blast
Posts: 5276
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: High positivity rate - why?

Post by jimbob » Mon Jan 10, 2022 3:29 pm

shpalman wrote:
Tue Dec 28, 2021 4:10 pm
It occurred to me I can check the preview csv as compared to the updated one I downloaded.

Preview:

Code: Select all

2021-12-22T11:00:00	3273144	9694628	6421484	698468	7.2	33367	226737	38197	16.8	645
2021-12-21T11:00:00	3267118	9661261	6394143	691474	7.2	34981	229644	36027	15.7	644
2021-12-20T11:00:00	3256991	9626280	6369289	685790	7.1	27032	221814	34052	15.4	643
2021-12-19T11:00:00	3250099	9599248	6349149	680999	7.1	29155	222165	33386	15	642
...
Downloaded just now:

Code: Select all

...
2021/12/22 11:00:00+00	3279936	9701419	6421483	698934	7.2	38503	233532	38666	16.6	645
2021/12/21 11:00:00+00	3268773	9662916	6394143	691580	7.2	35504	231303	36136	15.6	644
2021/12/20 11:00:00+00	3258123	9627412	6369289	685864	7.1	27031	222950	34129	15.3	643
2021/12/19 11:00:00+00	3251232	9600381	6349149	681073	7.1	29459	223302	33463	15	642
...
So yeah previous numbers get retrospectively updated.

The UK makes a clear distinction between cases by date reported (which almost never gets retrospectively corrected) and cases by specimen date (for which the most recent few days are incomplete). Recently I discovered that it was possible to download any dataset, even e.g. cases by specimen date or deaths by date of death as complled on different dates, so that you can reconstruct the evolution of the cases or deaths on a particular day as more and more data comes in.

Italy just reports a number every day (which almost never gets retrospectively corrected) which is usually taken to be the data from the previous day, but sometimes there are notes and corrections in the csv files. (I don't bother with the official Italian covid cases portal, just the vaccination one.)
The UK dashboard is good.

If you want demographic breakdowns, that's possible for England and its constituent parts but not the other nations.
Have you considered stupidity as an explanation

User avatar
shpalman
Princess POW
Posts: 8241
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond
Contact:

Re: High positivity rate - why?

Post by shpalman » Fri Jan 14, 2022 4:36 pm

The link for the csv of the COVID-19 Laboratory Testing Time Series seems to be there now, and up to date.
Ireland-14-Jan-2022.png
Ireland-14-Jan-2022.png (56.48 KiB) Viewed 1896 times
Now the positivity rate is really high (~50%) because the case rate is high (~20,000 per day).
having that swing is a necessary but not sufficient condition for it meaning a thing
@shpalman@mastodon.me.uk

Post Reply