To want to tear my hair out when people talk about the ‘R number’ increasing - IT DOESN’T MEAN WHAT YOU THINK IT MEANS!

[GinFling]

I am SO sick of everyone having a hot take on this virus - from friends and colleagues, and especially the media. Article after article about how the R value is increasing, nearly over 1, etc etc, and how this means we are heading for a second wave/disaster/the sky falling in. No, it is more complex and nuanced than that, and in fact it’s harder and harder to have a low R as the virus gets less prevalent. It is also hugely skewed by local outbreaks - such as in care homes and hospitals.

These two articles are quite helpful in understanding it:
www.bbc.co.uk/news/amp/health-52944037
unherd.com/2020/05/what-the-headline-covid-figures-dont-tell-you/

I wish everyone would just STOP using R to pretend to know what they’re talking about. We’re not all statisticians or virologists, for good reason.

[HesterShaw1]

The misunderstanding of the significance of the R number is a real thing. Peoole talk about it but don't know that the fewer infections the are at a population, the greater the effect just a few cases will have on the R of an area. Therefore the R number of the whole South West looks like it's high because of some cases at Weston hospital amongst staff. It's being used by the mayor of a town I live near (many miles from Weston) to say she "can't support" the reopening of shops and risks harming the small retailers even more. Because of some infections in a hospital well over 100 miles away.

[GinFling]

@Beatingthisthing exactly that - I was horrified to read a GP on Twitter stating that he believes the R number in his locality is above one, and should he therefore be telling patients to self-isolate again. It’s totally wrong - the R going up due to localised outbreaks in hospitals and care homes (whilst not ideal) - should not be informing public behaviour as a whole.

[itsgettingweird]

I agree R is misleading without understanding. I'm not going to pretend I understand but my belief was basically what beating said above.

I've watched things such as number of infections in testing. I've watched how that's a percentage of tests but it's been harder since they stopped publishing how many people were tested. I've watched them say estimates of daily infections from modelling. I've watched how it was 1:40 people to 1:500 to 1:1000 to now 1:1700. The latter 3 statistics being the last 3 weeks. I've watched rate of greatness compared to rate of decline.

I've also watched how an R of 0.8 means different things in different contexts. So outdoors and SD of 2m it means transmission will remain low. But indoors at no distance you could still infect 40% of the people present with the same R of 0.8.

But as a simple way of explaining it to people and getting them to comply it's a good indicator.

[GinFling]

@HesterShaw1 yes reading about this is what tipped me over the edge!

[TheCanterburyWhales]

Most of us who have been on the Covid science threads since January know what the R factor is thanks
No fucking idea what a "hot take" is though.
If you'd like to learn more about the science the Numbers and Graphs thread has some very intelligent people on it, who thus far, in almost 4 months have yet to resort to smug superiority and shouting at people.

[GinFling]

@itsgettingweird yes absolutely, the problem now is that the type of compliance required has changed. We now no longer want the general public to stay indoors - they need to be gradually getting back to ‘normality’ via shops, schools, etc. But the fear around the R value is keeping people from doing this, so schools aren’t going back, shops aren’t reopening, etc, when it’s actually perfectly safe to do so.

[itsgettingweird]

I agree. I have used statistics to tell me that during the height a trip to supermarket meant I was likely to have some contact (SD or a surface) with at least 2 infected people.

But now I think - I could go to a theatre and there is a small risk 1 person could be infected.

But that's what track and trace if for. That's why I think we should open up but have to leave details. Eg in a restaurant log time in and out, table number and phone details.

Spread will be fast in combined areas if there's an infected person.

Well that's my understanding. And my understanding is you shut this down by isolating everyone and cutting off transmission. The R rate doesn't really have any affect on this once the numbers are low like they currently are?

[GreyishDays]

[quote GinFling]@Beatingthisthing exactly that - I was horrified to read a GP on Twitter stating that he believes the R number in his locality is above one, and should he therefore be telling patients to self-isolate again. It’s totally wrong - the R going up due to localised outbreaks in hospitals and care homes (whilst not ideal) - should not be informing public behaviour as a whole.[/quote]
He is presumably taking it in conjunction with the number of cases though? I’m surprised you know more about epidemiology than a Dr.

[feelingverylazytoday]

Agree with you, OP.
But having said that, it's probably just better for you to stick to more fact/science based threads and forums, if it's making you feel irate.

[MedSchoolRat]

I'm an epidemiologist and I get them confused, Rt or R0, basic or effective... epidemiology is a lot bigger than R-numbers, though.

I don't think what OP is saying is unreasonable. If only 2 people have covid today then it doesn't matter if they each have 12 contacts that get sick, as long as none of the 24 contacts pass it on, or only pass it onto 0.1 people each. So R number could be 12 this week and 0 next week... that's where R0 & Rt difference matters. R0 is the longer trend, it's R0 that mustn't accelerate away. Rt not so worrisome. Journalists keep trying to publish Rt as 'the R number' but actually it's R0 we need to know about. Both imperfectly measured.

Constantly we come back to the bigger conversation: are we trying to stop all covid-19 or is there a level of circulating covid & consequences that we can live with (say 5 deaths/week). If you love one of those 5 people, then 5 deaths is too much covid. Or you could say that we routinely live with 5 deaths/week from influenza without shutting economy & human relationships down, so 5 is tolerable. I wish this grown-up conversation would happen.

[BuzzShitbagBobbly]

Since covid 19 we're all doctors and scientists. We are now all experts on diseases and viruses...and I include myself.

Makes a change, over the winter we were all hydrologists and flood management experts.

Amateurs. Everyone on my social media was all those things AND world leading experts on Brexit.

[SockYarn]

I understand the basics - R above 1 the virus is growing, R of 1 it's remaining the same, below 1 and it's reducing. Because at a R of 1, every person will infect one other.

In Scotland our R rate is something between 0.6 and 0.8 according to scientists and Ms Sturgeon but it's a work of fiction because it includes settings like care homes or hospitals where one person is perhaps infecting a dozen others. In the community, rates could be much lower.

I think it would be more useful to look at other indicators such as the number of people admitted to hospital each day to get a true picture of what's happening.

[mocktail]

I completely agree. I'd rather be in an area with very low infections but an R rate of 1 than in an area of high infections but an R rate of 0.9.

If 1 person in the entire country had coronavirus then passed it on to 2 other people then the national R rate for the entire UK would be 2. R rate is only meaningful when looked at in combination with other factors.

[GrandAltogetherSo]

According to the news today, Beijing had weeks of no cases and have now returned to Lockdown after a sudden spike in cases that emanated from a market.

I guess if they follow your rules, then they didn’t need to lockdown because although the R factor is above 1, the virus is contained in a small-ish area and they should carry on, business as usual?

[fadedout]

Ewwww 'hot take'

[MedSchoolRat]

ps: the other problem with covid is the lag from exposure to detection (ascertainment). If someone isn't diagnosed until 10 days after their exposure, then the chains of exposure may have grown a lot in the mean time. You need a few cycles (lots of 10 days) to be sure what you're seeing. For a respiratory disease easily passed on, 30 day windows back into the past means things could have accelerated away badly.

I thought the genomic data were now suggesting that there was a hute influx of unique new lineages into Britain in late Feb/early March. There was only a few lines modestly circulating in February, in contrast. It literally took 20-30 days to get a full view of how much more covid was now circulating due to the new introductions in early March (esp. from Spain, Italy), and hence Lockdown happened near end of March.

[DontTouchTheMoustache]

You cant get frustrated about people misunderstanding something as a simple rating system when the government have explained it as a simple rating system.
People have taken what they have been told at face value and people are scared.

[Northernsoullover]

You need to give people more credit. I study this for my university degree and while I have a lot to learn I do know enough to present a confident view. YABU. Very.

[Eckhart]

Don't worry. You'll have torn it all out soon, given that so many people aren't doing things the way you want.

Then your problem will be solved.

[mrsBtheparker]

Since covid 19 we're all doctors and scientists. We are now all experts on diseases and viruses...and I include myself.

It makes a change from everyone knowing all about education because they went to school and they've bred!

[scaevola]

Number depends on three things - two of which are characteristics of the virus (which we can do nothing about). It is also based on an entirely vulnerable population

Characteristics of the virus:

1. how long an infected person is contagious for, including before symptoms show
   
2. the secondary attack rate - how many of an infected person's contacts will catch it
   
   Influenced by people:
   
3. how many people an infected person come in contact with
   
   So yes, a cluster in a care home, prison or other setting when many people are in contact, changes the Rnumber. But unless the workers in these settings are live onsite; the chance of them taking it home, the shops, anywhere where they might have contact with other people also increases

[HesterShaw1]

In my head I always compare it to the situation I used to find myself in with Y6 classes (bearing in mind I left teaching 15 years ago). If I taught in a year group of 70 which contained a couple of children who didn't do their SATs because they had statements of SEN, that would have far less of an effect on our overall results than the school who had 9 Y6 children,8 of whom got level 4 or 5s apart from the one child with a statement who didn't sit the tests (as I ssus I left teaching AGES ago so my terminology is way out of date). However looking g at the basic figures and nothing else the small school's results look worse.

As an aside is it possible to have even one discussion about this without people getting defensive?

[Iloveappleproducts]

I'm not everyone and I'm not a scientist but I have an understanding of the R number.
That's because I can read and take in information
Your post is patronising
How do you know what 'everyone' understands?

[mocktail]

As an aside is it possible to have even one discussion about this without people getting defensive?

^^ agree!

[EasyAndy101]

Science in the press works like this:

Limitations in detail render the whole thing fucking pointless, a 400 page study gets reduced to a few inches in a newspaper column or a few minutes on TV. That reduction being done by people who themselves did not understand the data and also knew that most of the people that read the column won't understand it either