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.

It looks from further detail to be much better news from a public health perspective, because it doesn't have many asymptomatic carriers with the low followed by high dose (a massive risk of these vaccines), as well as going up to 90% effectiveness: www.bbc.co.uk/news/health-55040635.
The storage and lower cost makes it a far more likely candidate for global distribution; if that happens it would be fantastic economically for the UK.

[quote Firefliess]@Augustbreeze You can't simply add together the bars in that chart for all the types of teacher. They're proportions, not simple numbers. What you want is a kind of weighted average. You'd need to add together all the teachers who got Covid in all three groups, and divide by the total numbers of all three types of teachers (ie the the separate denominators) which I don't think is possible from the data available in the publication. The suggestion made in the Twitter feed is instead to focus just on the "teachers of unknown type" because the error bar on that group is smaller, which means that it's a much larger group than the other two types of teacher (ie most teachers did not specify whether they were primarily or secondary teachers). [/quote]
I thought that too but on another thread a poster pointed me here - www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/datasets/coronaviruscovid19infectionsurveydata/2020 - the update for 13 Nov includes (tab 1k) figures for the actual numbers so you can do this:

% testing positive for COVID-19	95% confidence interval			
	Lower	Upper	Individuals testing positive for COVID-19	Total individuals in sample

Carehome (resident-facing) 0.00 0.00 0.60 * 611
Education support staff 0.45 0.22 0.82 10 2227
Health care (non patient-facing) 0.51 0.23 0.97 9 1759
Health care (patient-facing) 0.37 0.25 0.54 28 7546
Nursery/preschool teacher 0.45 0.05 1.61 * 445
Other professions 0.44 0.39 0.49 286 65047
Primary school teacher 0.23 0.07 0.53 5 2190
Protective service occupations 0.52 0.25 0.95 10 1936
Secondary school/6th form teacher 0.38 0.15 0.78 7 1852
Teacher of unknown type 0.51 0.36 0.69 39 7714
University lecturer 0.41 0.08 1.18 3 739
Other key worker 0.40 0.29 0.53 47 11852

Sorry, the formatting of that is nonsense, but you can see it in an actually usable format at the link

100 million doses at 1.5 doses per person also means enough to vaccinate (practically) every single person in the UK

I seriously hope we vaccinate at risk people not in the UK before we consider dosing very low risk people in the UK.

@sirfredfredgeorge

100 million doses at 1.5 doses per person also means enough to vaccinate (practically) every single person in the UK

I seriously hope we vaccinate at risk people not in the UK before we consider dosing very low risk people in the UK.

That won't be a UK decision.... there is nothing we can do to influence how quickly other countries can vaccinate their population

I seriously hope we vaccinate at risk people not in the UK before we consider dosing very low risk people in the UK.
Wow. I really don't (barring Ireland, as is traditional they can share).

1. You're taking about a couple of billion people globally (all healthcare, all over 50, all with certain health conditions), we'd be on lockdown for at least a year longer
2. British taxpayer money and British scientist expertise created a vaccine; we've had enough critique that the UK has had a high number of deaths, how about we take the benefit of how the UK works as well as downside?
3. Many countries, outside a select few in Europe and the Americas, aren't declaring huge case and death levels, so why should their citizens be protected before UK citizens who are at risk due to high case levels? The UK will remain at risk because of our position as a transport and business hub; the same factors causing early infection seeding can keep causing it with mutated strains again and again.
4. EU, USA, Australia, Japan etc all had the option to buy whatever doses they saw fit, developed countries don't have extra need.
5. The UK is contributing just the same as others to the global fund for vaccines as a charitable offering. The few million extra UK is a drop in the ocean in terms of global need.

I've just been looking at that same article @clare. With a bit of maths using the info they've given, I think that means that 2 of the 1371 people who got the first half dose caught Covid, compared to 24 in their placebo group (the other 1371 in their trial) And in the group given two full doses about 28 of 4448 caught it and 77 of 4448 given the placebo group caught it. Those figures seem to add up to the figures we've been given - 90% effectiveness for the small dose trial and 62% for the 2 full doses, giving 70% (ish) between them. (I think the reason my numbers don't quite add up is that the people have been in the trial for different lengths of time and they weight the data to allow for that)

But I've now reached the limit of my maths skills and cannot work out the confidence intervals for the smaller trial alone. Can anyone else do this? Ie, at a 95% confidence interval, how effective is the vaccine if given with small first dose? Intuitively I'm thinking around 80-95%? But I'd prefer maths to instinct!

@sirfredfredgeorge

100 million doses at 1.5 doses per person also means enough to vaccinate (practically) every single person in the UK

I seriously hope we vaccinate at risk people not in the UK before we consider dosing very low risk people in the UK.

But as we are always told we have one of the worse death rates in the world. So this is one occasion when I think we should put our own country men and women first. Other countries have suffered less than us ( for a variety of reasons) and some countries have no Covid community transmission at all (NZ springs to mind) so they are less in need of a speedy vaccine roll out than us in the U.K./ Europe.

We need to work out the best thing for the whole world. Covax is trying to do this. The UK does have a say in this.

www.who.int/initiatives/act-accelerator/covax

With a fast-moving pandemic, no one is safe, unless everyone is safe

Thanks for the link to that data @hardbackwriter. You can indeed see the raw data. When you add the three types of teacher together, the rates for teachers comes out a 0.43% as compared to 0.44 for other professions, and 0.43 for all employed adults. The confidence bars on the chart they published would have been shorter if they'd done that but the end conclusion that there's no evidence from this survey that teachers overall are at higher risk does in fact remain.

I'm not sure why the mathematician on the Twitter feed didn't do this for herself (maybe she didn't find the data?!)

The dates here also appear to be the same for all groups (2 September to 16 October), so I'm not sure why she thought there was a problem with that? Maybe I should be more skeptical of people on twitter, even if they are maths professors!

we'd be on lockdown for at least a year longer

So? Why is it morally better to let a non Brit die, than a Brit?

It's nothing about cost, it's about SUPPLY, for me personally it is utterly morally wrong for me to have a vaccine than a vulnerable person, whatever their nationality.

Obviously this doesn't address "herd immunity", but because the vaccine has a 90% chance of preventing a vulnerable person becoming infected and me or not being vaccinated does almost nothing to achieve it (we just need R well below 1 for awhile, so need to go for very high levels of vaccination)

10000 doses on vulnerable people in Spain might save 6000 infections and 60 deaths, 10000 doses on 20 year old healthy brits might save 6000 infections and 0.05 deaths. I believe it's wrong to use up the scarce resource for the second of those.

but let's not get into it on the data thread!

@Firefliess the dates matter because the length of time students were back at school influences transmission rates in schools. And the 17 Oct is when secondary infection rates started to really rise.

Surely there are 4 groups of teachers there, not 3: nursery, primary, secondary and unknown?

Thanks for @Hardbackwriter for posting the source.

The original ONS report defines Other Workers but it's not defined here, will have to go back and check.

77 of 4448

Again, an absolutely stunning result - what's the explanation this time, it's in the UK, the UK has had over 5% of 18-65 year olds infected since the vaccine started according to the ONS, yet the people who got a placebo only 0.17% Even if we required symptoms to be tested in the vaccine study, then it's only 1/5th of what we'd expect for the general population.

Is the placebo working that well?

As I said, someone else pointed me to it, after I confidently asserted that it wasn't available! I suspect the twitter poster thought the same - it really isn't easy to find if you don't know it's there.

Also, just to say that this is the first time I've ever posted on one of the data threads because it's the first time that I've felt sure I had something of value to add! But I've been reading them on and off since about April - I think they're the best threads on MN

@sirfredfredgeorge

we'd be on lockdown for at least a year longer

So? Why is it morally better to let a non Brit die, than a Brit?

It's nothing about cost, it's about SUPPLY, for me personally it is utterly morally wrong for me to have a vaccine than a vulnerable person, whatever their nationality.

Obviously this doesn't address "herd immunity", but because the vaccine has a 90% chance of preventing a vulnerable person becoming infected and me or not being vaccinated does almost nothing to achieve it (we just need R well below 1 for awhile, so need to go for very high levels of vaccination)

10000 doses on vulnerable people in Spain might save 6000 infections and 60 deaths, 10000 doses on 20 year old healthy brits might save 6000 infections and 0.05 deaths. I believe it's wrong to use up the scarce resource for the second of those.

but let's not get into it on the data thread!

Spain will have access to all the same vaccines via the European vaccine initiative. I imagine Spanish people will also be vaccinated pretty quickly

@sirfred. I don't think the Oxford trial participants were all in the UK. Some were in Brazil and (I think) the US. They went abroad when it looked as if our rates were falling too low for them. But I agree, it is really puzzling that the rates are so low in the placebo group. There have been about 1m UK positive Covid cases (diagnosed) since August. That's more than 1% of the population. And the other countries they choose have had similar rates. Are there big delays between people catching Covid and the data feeding into the study? Are the types of people who volunteer for trials really good at social distancing? Did their efforts to target high risk occupations mean they accidentally recruited large numbers who's already had Covid (maybe with t cell immunity rather than antibiotics, which I assume they checked for, but maybe not?) I don't know the answers but would like to see someone asking these questions!

@cathyandclare

The difference in results with different dose regimens is interesting. There are fewer in the half dose/full dose group but the results are statistically significant.

Does anyone have a view on which dose they are likely to go with?

A smaller group but better results? Not sure what the thresholds are (no idea really)

@Augustbreeze - the number of nursery teachers in the study was tiny so the number of positive tests they had so small it has been surpressed (ie it is less than 5, and quite possibly zero) So it's not possible to add them in to a teacher total (and would make no significant difference to the findings if you did) You can include the educational support staff, but their rates are also pretty much identical to those of teachers, so it makes no difference to the findings.

Yes rates have increased in schools since mid October, so it would be good to redo this analysis soon. But that doesn't invalidate it as preliminary analysis based on the early data available.

Why would they not go with the dose that gets the best results?

@TheCountessofFitzdotterel

Why would they not go with the dose that gets the best results?

Maybe they will I guess any group number they choose they must rely upon as a result otherwise why choose it (wondered if a group size was considered too small)

I'm sure they'll go with the lower dose. Firstly, the data suggests it is more effective. And secondly, it only requires 1.5 doses per person so they'll be able to manufacture enough for more people more quickly.

In which case we could see a vaccine with up to 90% and not at very low temps so avoiding logistical issues

Yes. It’s wonderful!

The group vaccinated with initial half dose was only a smaller subsidiary group so it needs to be looked at in more detail before decision made

That's what the head of the Oxford vaccine group said in his bbc interview this morning