Data, Stats and Daily Numbers started 17th March

[boys3]

UK govt pressers Slides & data www.gov.uk/government/collections/slides-and-datasets-to-accompany-coronavirus-press-conferences#history
R estimates UK & English regions www.gov.uk/guidance/the-r-number-in-the-uk
Imperial UK weekly LAs, cases / 100k, table, map, hotspots statistics Attendance explore-education-statistics. service.gov.uk/find-statistics/attendance-in-education-and-early-years-settings-during-the-coronavirus-covid-19-outbreak
NHS England Hospital activity www.england.nhs.uk/statistics/statistical-work-areas/covid-19-hospital-activity/
NHs England Daily deaths www.england.nhs.uk/statistics/statistical-work-areas/covid-19-daily-deaths/
Cases Tracker England Local Government lginform.local.gov.uk/reports/view/lga-research/covid-19-case-tracker
ONS MSAO Map English deaths www.england.nhs.uk/statistics/statistical-work-areas/covid-19-daily-deaths/
CovidMessenger live update by council district in England www.covidmessenger.com/
Scot gov Daily data www.gov.scot/publications/coronavirus-covid-19-daily-data-for-scotland/
Scotland TravellingTabby LAs, care homes, hospitals, tests, t&t www.travellingtabby.com/scotland-coronavirus-tracker/
PH Wales LAs, tests, ONS deaths Dashboard app.powerbi.com/view?r=eyJrIjoiZGYxNjYzNmUtOTlmZS00ODAxLWE1YTEtMjA0NjZhMzlmN2JmIiwidCI6IjljOWEzMGRlLWQ4ZDctNGFhNC05NjAwLTRiZTc2MjVmZjZjNSIsImMiOjh9
ICNRC Intensive Care National Audit & Research reports www.icnarc.org/Our-Audit/Audits/Cmp/Reports
NHS t&t England & UK testing Weekly stats www.gov.uk/government/collections/nhs-test-and-trace-statistics-england-weekly-reports
PHE Surveillance reports & LA Local Watchlist Maps by LSOA www.gov.uk/government/collections/nhs-test-and-trace-statistics-england-weekly-reports
ONS England infection surveillance report each Friday www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/coronaviruscovid19infectionsurveypilot/previousReleases
Datasets for ONS surveillance reports www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/datasets/coronaviruscovid19infectionsurveydata/2020
ONS Roundup deaths, infections & economic reports www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/articles/coronaviruscovid19roundup/2020-03-26
Zoe Uk data covid.joinzoe.com/data#interactive-map
ECDC rolling 14-day incidence EEA & UK read https_www.ecdc.europa.eu/?url=https%3A%2F%2Fwww.ecdc.europa.eu%2Fen%2Fcases-2019-ncov-eueea
Worldometer UK page www.worldometers.info/coronavirus/country/uk/
Our World in Data GB test positivity etc, DIY country graphs ourworldindata.org/coronavirus/country/united-kingdom?country=~GBR
FT DIY graphs compare deaths, cases, raw / million pop ig.ft.com/coronavirus-chart/?areas=gbr&areas=fra&areas=esp&areas=ita&areas=deu&areas=swe&areasRegional=usny&areasRegional=usnj&byDate=1&cumulative=1&logScale=1&per100K=1&values=deaths
Alama Personal COVID risk assessment alama.org.uk/covid-19-medical-risk-assessment/
Local Mobility Reports for countries www.google.com/covid19/mobility/
UK Highstreet Tracker for cities & large towns Footfall, spend index, workers, visitors, economic recovery www.centreforcities.org/data/high-streets-recovery-tracker/

⏭ Our STUDIES Corner ⏮www.mumsnet.com/Talk/coronavirus/3869571-Studies-corner?msgid=99913434

We welcome factual, data driven and analytical contributions
Please try to keep discussion focused on these

Ha! How can he resist?

@JanFebAnyMonth

https://mobile.twitter.com/deeksj/status/1372506978661109760?s=20&fbclid=IwAR2lEfggu55Pz9czoY5VpCae6RI6h8UHDMHCdr9M_NtV2wTHyGl3fBhfdzE

John Deeks' view on the in-school LFT results. Let's hope Michael Mina responds?

It's an interesting discussion. But why does he think that the true rate should be necessarily be close to 0.5%? That's the overall prevalence across society (a couple of weeks ago) including people with symptoms who know that have Covid Nobody involved with the school testing programme had symptoms, nor would they have been contacts of known cases (who are instructed to stay home and not go to school/work) So if you test an entirely well population, obviously the proportion who in fact have Covid is going to be a lot lower than a cross section of the whole population, including those with symptoms, who've had positive tests or are in isolation as contacts.

Yes, also:

1. Many of the most exposed are kids of key worker parents have been going to school throughout, so more likely to have already had covid than others rejoining school now
2. Those whose family are covid deniers / unwilling to follow rules because it doesn't affect them / generally not following rules - would be very unlikely to consent to LFD testing
3. Children have always had a scale of infection whereby only those near adult have neared adult infection rates and there is some drop in infection rates, by age as borne out by various surveys (putting aside responder bias there); lockdown would have suppressed those cases further due to point 1
4. So what if it only finds 100 cases, if that saves only say 800 kids and 50 teachers from getting infected next, plus 850 more families making up 2500 people... By that stage of thousnds we are already way past the likely numerical point where we reach enough to catch vulnerable needing hospitalisation. The point of why LFDs have benefit despite false negatives is that they catch many of the most infectious, who are most likely to be huge superspreaders.

Looked today at the rate of LFD tests in each council area; as in the the number of tests per 100,000 population as opposed to positivity rates, with this for the week up to Monday just gone - so a three day lag. I do wonder about how widespread the enthusiasm for continuing to test at home; and, next step reporting the results as well will be. Or indeed how the take up rate has varied in secondaries / sixth forms across the country.

The graph plots the latest rate of LFD tests, so up to Monday 15th and over seven days when a lot of in school testing would have been taking place against a council areas overall deprivation score. This is a somewhat crude comparison demographic profile is not considered, the test rate is nothing more complex that total number of tests / total area population.

I thought it may be worth repeating the same time next week to see if the pattern changes. I suspect given the self-isolation cost for many the picture may evolve further.

Oh that's really interesting boys! So, some correlation between high levels of deprivation and low levels of testing, but not a great deal.

(Non-mathematician here)

@JanFebAnyMonth

Oh that's really interesting boys! So, some correlation between high levels of deprivation and low levels of testing, but not a great deal.

Relative levels of deprivation are higher for proportionally younger populations. The link is likely much stronger if one axis had population 11-18 divided by number of tests. (Hint hint hint)

That is interesting @boys3. Lower rates of LFTs in more deprived areas, despite the fact that those areas have higher proportions of young people and keyworkers, who you'd assume would be the most likely to be being tested.

Cross post with @Mrex! Yes, it would be good to do the rate of 11-18s, as most LFTs have been secondary aged. (If you have the time @boys3!)

no pre-empting my weekend project @mrex

More seriously stripping out test numbers by specific age bands would provide a more clear cut view. Hence given the limitations of what is readily available looking at the same data set a week forward may be an option.

Lower rates of LFTs in more deprived areas

It's index of multiple deprivation, so areas with lots of unhealthy older people are in also highly deprived despite not having that many kids, so whilst many of them will have higher numbers of younger kids. There will also be many with few, and hence few tests.

Exactly the ratios, no idea, but I guess we'll find out at the weekend when boys has done all the work for us.

.

@sirfredfredgeorge

Lower rates of LFTs in more deprived areas

It's index of multiple deprivation, so areas with lots of unhealthy older people are in also highly deprived despite not having that many kids, so whilst many of them will have higher numbers of younger kids. There will also be many with few, and hence few tests.

Exactly the ratios, no idea, but I guess we'll find out at the weekend when boys has done all the work for us.

On average, the local authorities in England that rank higher on the IMD (ie most deprived) have more younger people in them. (Older people have had more time to amass housing equity, so typically move to nicer, rural areas. Young people - and immigrants in particular - prioritise jobs and housing they can afford over nice areas) And yes, there's your weekend's work cut out @boys

And yes, there's your weekend's work cut out

so swiftly moving away from deprivation a small majority of councils had falling rates over the last seven days. As numbers in many area start to get smaller percentages can get a bit meaningless, so the upward or downward movement on the y axis is the absolute change in the rate per 100,000; so if council "x" was 8 per 100,000 in the prior 7 days and is currently 20 for the most recent 7 days its increase is 12 on the graph as opposed to 150%. Longwinded way of saying those with an already relatively higher rate have been more likely to have increased.

If anyone is sad enough to try to count the dots Derbyshire Dales is not shown as the size of its drop ruins the axis scale. Its gone from just under 180 per 100,000 to now 35 per 100,000. Appears prison outbreak well and truly under control; clearly good news.

Enough of this data stuff I need to go and worry about whether poor young Billy is about to be blocked.

@boys3

no pre-empting my weekend project *@mrex*

More seriously stripping out test numbers by specific age bands would provide a more clear cut view. Hence given the limitations of what is readily available looking at the same data set a week forward may be an option.

Yes, removing average January/ February weekly LFT totals from the LFT totals would be great. Interesting for other LFTs to do a "previous" comparison of compliance too, maybe... though there are so many variables it could get tricky.

Just delurking to let you know of this study published in the BMJ yesterday: www.bmj.com/content/372/bmj.n628
It shows the risks of infection and outcomes (hospitalisations and deaths) of those living with children in England. Basically it quantifies the (small) increased risk to adults when the schools were open in wave 2 compared with wave 1 when everything was shut.

Really interesting @XJen, thanks.

@XJen

Just delurking to let you know of this study published in the BMJ yesterday: www.bmj.com/content/372/bmj.n628 It shows the risks of infection and outcomes (hospitalisations and deaths) of those living with children in England. Basically it quantifies the (small) increased risk to adults when the schools were open in wave 2 compared with wave 1 when everything was shut.

Thanks for that. My reading of it (skimming!) seem to be that parents are at slightly higher rate of catching Covid when the schools are open (22% increase if you have secondary aged children; 6% for primary) But parents are a healthier lot than non-parents to start with (unsurprising - people with severe health problems and learning disabilities are less likely to have children), so at a lower risk of hospitalisation regardless.

The 6-22% increased risk of infection seems broadly online with the increased risks that teachers see too, from other studies - a measurable increase, but small compared with some other occupational groups. Parents of course see a lot less children than teachers, but have much closer contact with them, including unavoidable contact when they're sick. So it seems plausible to be. Roll on the vaccine for the 40 something parents of secondary aged children!

That's very interesting.They did actually look at the increased risk to parents who had been shielding of having a child going to secondary school. Looking at their figures, it does look like there was a significant increase in risk (Hazard ratio 1.46 with confidence intervals0.95 -2.24)

So this was a very interesting, very positive looking finding. Very interesting to me, as one of the aforesaid shielding parents.

How curious that this was not further discussed in the conclusion.

They also did not mention the numbers of parents thought to be shielding, and there appears to have been little attempt to find out whether these families were actually shielding or not. This would have a major effect on the supposed risk.

I would assume, from the failure to pursue this, that the researchers were hoping/ expecting to knock out a result that fits with the political narrative of schools not being much of a risk to parents overall (and not paying much attention to the exceptions).

Bit disappointed in Ben Goldacre really.

That's very interesting. They did actually look at the increased risk to parents who had been shielding of having a child going to secondary school. Looking at their figures, it does look like there was a significant increase in risk

I agree, 22% is more than 1 in 5 which is a significant rise in risk. I wonder if you have more than one child does it increase further? I suspect so.
It would also be interesting to see if the lower primary increase is due to the children being younger or that the classes tend to be much more defined so the bubbles are smaller (ds is in a bubble of 300, dd2 in a different bubble of 200-that must be a higher risk than my friend whose year 7 stays within their form of 32 the whole time)

Sorry, the risk was broken down further.

For "probably" shielding parents, with secondary age children, increased hazard ratios were

hospital admission 1.71 (1.40-2.09)
ICU 1.58 (1.08-2.33)
death 1.25(0.67 -2.34)

I agree, 22% is more than 1 in 5 which is a significant rise in risk

It's not absolute though. I am not a great statistician, so probably won't explain well, but the increase of 22% does NOT mean 22 people out of one hundred. It means, for example, if the background risk is 1 in 100, a 22% increase, only moves that to 1.22 per hundred. Am I right?

It wasn't clear to me which risk they were talking about, to be honest.

As a shielding person, my risk is higher than Ms Not-a-shielding-Parent.

If my risk for some outcome is 1.71 times higher because my kid is going to secondary school, are we comparing my current risk with

• my risk if I were shielding and didn't have kids
• the risk of Ms Not-a-shielding-Parent also with kids at school
• average risk for my age

-average risk for the country

It wasn't immediately clear to me on looking through their methodology how they had considered which risk for the comparator.

There is also a major confounder with occupation which they admit they haven't really tackled.

@BigWoollyJumpers

I agree, 22% is more than 1 in 5 which is a significant rise in risk

It's not absolute though. I am not a great statistician, so probably won't explain well, but the increase of 22% does NOT mean 22 people out of one hundred. It means, for example, if the background risk is 1 in 100, a 22% increase, only moves that to 1.22 per hundred. Am I right?

Yes you're right.

So out of 122 parents with Covid during the November lockdown, only 100 of them would have had Covid if the schools had been shut.

Interesting that they looked at shielding parents - you'd expect their increased risk to be a lot higher because catching Covid via a family member is the main risk that shielded people encounter. I suspect the reason they've not made more of the findings though is that the numbers of shielding parents in their data is low, and so the error margins are high - ie they can't be very confident that the increased risk they see didn't just occur by chance

@herecomesthsun

It wasn't clear to me which risk they were talking about, to be honest.

As a shielding person, my risk is higher than Ms Not-a-shielding-Parent.

If my risk for some outcome is 1.71 times higher because my kid is going to secondary school, are we comparing my current risk with

• my risk if I were shielding and didn't have kids
• the risk of Ms Not-a-shielding-Parent also with kids at school
• average risk for my age

-average risk for the country

It wasn't immediately clear to me on looking through their methodology how they had considered which risk for the comparator.

There is also a major confounder with occupation which they admit they haven't really tackled.

They're comparing you with the risk you encountered in wave 1 when the schools were shut.