Data and analysis thread, started 12 November

[NoGoodPunsLeft]

Previous thread here:

Data and Analysis Thread, started Oct 29 www.mumsnet.com/Talk/coronavirus/4064113-Data-and-Analysis-Thread-started-Oct-29

Regular lurker but I frequent poster, didn't want to lose the threads.

Schools close if they don't have enough staff to ensure safety of all on site. So if staff are self isolating or taking time to look after young children, that can also be a problem. In my son's school it's not yet been teaching staff who have been affected, so despite 4 bubbles out, they're ok.

Looking at today's data and the surveillance report (thanks for the link, had forgotten they'd combined it with flu) I think it's maybe wrong to try to focus on the national levels. There is a big decline in the North West, and still increases in many other regions, especially in the south. I would guess that's a slow impact from the tier 3 restrictions, but still not much impact from the nationwide lockdown as yet.

Just in case this not posted yet - Hull daily mail - schools on brink of collapse, 15,000 students and teachers absent from school, 64% of students isolating or shielding - more data in article.

www.hulldailymail.co.uk/news/hull-east-yorkshire-news/hull-schools-brink-of-collapse-4716574

Graph from @RP131 on infection in Liverpool. What do you think? There's a steady fall, but it seems to predate measures. Tier 3 doesn't look like it's changed the curve much. The moonshot testing has slowed the speed of decrease, presumably due to asymptomatic diagnosis.

@IloveJKRowling yet you won't see that in any national media... I remember when the daily mail were raging about a handful of cases in meat processing factories!

Yes you're right @cathy, the fall in Liverpool does seem to predate the introduction of Tier 3. Though didn't they have some significant restrictions before then, as part of the watch list type stuff?

Yes, I think Liverpool’s case proves that Tier 2 restrictions combined with public awareness (that may be driving voluntary changes to people’s behaviour) can be effective.

Cathy, that looks cautiously positive.

Fireflies, I agree. Things have been spiralling in the NE and Hull as below.

One area falling cancels out a rise in another area. So nationally good, locally some really bad spots that have huge implications. Especially as most of those areas are areas of high deprivation already.

I keep thinking about a recent report about how areas that had high bombing in the blitz are still suffering from higher rates of poverty. I fear a slightly similar longer term impact with Covid on some areas.

Irrespective of schools being a potential issue in terms of transmission rates, the impact of rising rates fully closing schools and bubbles at a local level is a huge issue. I know of a number of schools that are closing for two weeks locally and I'm not Hull.

And, of course, there is also a question of acquired immunity:

assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/935238/potential-trajectories-covid-190next-6-months-s0848-291020.pdf

I have already posted this link. It does show that even conservative scientific advisors (SAGE) are reluctantly accepting that (i) some acquired immunity does exist and (ii) when combined with other social distancing measures, it can help to drive R from slightly over 1 to under 1 which helps get transmissions under control.

I have a question that I don’t think has been covered.

I keep track of the daily government figures and the ZOE app. I’ve seen that there’s been discussion on here that ZOE may not be ultra reliable but what do people make of the gap closing between reported positive cases and estimated new cases from ZOE? Is that more likely to point towards testing capturing more cases or ZOE’s estimates not been as useful as they were at one point?

Zoe's most recent estimates at the national and regional level are quite close to ONS (adjusting for 20% of asymptomatic cases) so I personally think the aggregate numbers are reasonably accurate. I would not use Zoe to check levels of infection in my London borough but I do look at it daily for general trends.

As a result, I actually think we are now picking up a higher proportion of cases than we used to.

Zoe's user numbers keep increasing - roughly a million more (4.4 million now) in about seven months I'd say - so it's becoming more and more representative I guess?

Isn't the 4.4 million the world wide number for the app, not the UK number, it seems surprisingly high?

Surely the vast majority of their users are UK??

Thanks @Augustbreeze and @wintertravel1980.

With the disclaimer ZOE is estimates; it can only be a good thing we’re formalising a good proportion of their estimate with testing. That gives me some comfort there has been some progress made... not that the bar was set very high in the first wave!

4.4m.are UK users, or at least claim to be: covid.joinzoe.com/.

wintertravel1980 It’s a shame that SAGE have used inaccurate data for Florida in the link you posted above. That is death by date of reporting not actual date of death which show a higher peak and more dramatic decline. Doesn’t exactly fill one with confidence if they are basing what they think will happen in the UK on that. This is the date of death chart.

Is Zoe daily contributors though? I'm afraid I stopped using it in sept for a variety of reasons.

I think Zoe's been over 4 million for ages, in fact I'm not sure it's changed much since I joined back in April/May time.

Our area on Zoe has approximately 5 times the number it had this time last week. Our area is going up quickly on official figures too.

Good point @NeurotrashWarrior, wonder if they have a (published) figure for contributors by day.

It was around 3.5 m when I joined in late March, I'm sure.

I don’t know about the total daily contributor rate but they do publish how many have contributed each day by area on their daily UK heat map. There are occasionally areas greyed out because there haven’t been enough contributors the previous day so they’re clearly following a minimum number to make their estimates.

ZOE is pretty much part of my daily morning routine now and I record for all three of us in the house. I agree the local estimations aren’t great, I can’t quite figure out how they can vary so much by day (up and down) and whether or not they’re comparing the previous day or corresponding day last week... the local numbers can be over a very wide spectrum.

I don't think they say, and I think they should. They are also a little hazy in the distinction between how many contribute and how many records they are tracking; most people using it are probably like me and completing it for the family, so they get multiple person-results per daily user.
By the way, they put R at 0.9 now, which is great:
covid.joinzoe.com/post/covid-declining-across-the-uk.

The difficulty is that things vary so much by region that the R value for England doesn't really tell us very much I don't think. I'm in the South East and cases seem to have stabilised but aren't dropping. Levels are low and so if that can be maintained then it might be enough to limp through until the vaccine here.

Other areas are rising despite lockdown which is obviously a problem and the question is why? Is it non compliance or something else?

Anecdotally in my area people are still socialising outside, lots of families meeting at woods/playgrounds/parks etc. We have complied with the lockdown but I can't say I judge people for meeting others at this point.

From what I can see, it’s not even variance by region. You can’t just look at the whole North East or South West. If you look at the map, the areas affected are much more localised than that.

I am watching various areas (local authorities) local to me and some are doing remarkably well and have cases on a downwards trajectory whereas others have case numbers still shooting upwards at quite an alarming rate.