New very large study suggesting over 75% of people with the virus are asymptomatic - and over 80% without ‘core’ symptoms - much higher than earlier studies, but they were mainly done in environments like cruise ships, where more older people were probably more likely to develop symptoms.
If the vast majority of people with the virus have no or few symptoms, it might already be much more endemic - and much harder to control - than previously thought.
To me it was already obvious that testing had to include those who asymptomatic as well. It is common sense.
This is a failure of the ‘world beating’ testing system and now with Roche saying that they are having a problem in supplying swabs for testing, it can only get worse.
I am at risk because I am over 70 and asthmatic. I do not feel that I am being discriminated against by shielding myself, so those social scientists need a kick up the backside and told to get real.
Again it makes sense if you are at risk to avoid situations where you might catch the virus.
I know that we are lucky and live in a beautiful part of the country with space around us and not in a small flat with no access to outside space, but I wouldn’t knowingly put myself at risk by walking down a street late at night in a high crime area either.
2 Likes
In this case you really have to be careful about false positives.
The ONS are giving the number of infections at about 0.2-0.25% at present.
When considering any test you think in terms of what are known as the sensitivity and specificity.
Sensitivity is what %age of patients with the condition will be picked up, if 9 out of 10 then we say the sensitivity is 90% and that the false negative rate is 10%
Specificity is how many without the condition are correctly identified - it tells us about the false positive rate.
Let us suppose a test which is 99% sensitive and 99.8% specific (0.2% false positive) - this is a test which is performing rather better than most estimates for Covid tests.
Suppose that the true incidence is 0.05% and we test 10,000 people
5 people have the disease, odds are 5 will test positive (5x.99 = 4.95).
What of the rest, who do NOT have the disease, well (10,000-5)x0.2% = 20 false positive tests.
So we have 5 positive tests who have the disease
And 20 positive tests that do not have the disease
A total of 25 positive tests - 0.25%
Naively you might suppose that 80% (20/25) were “asymptomatic carriers”
You would be wrong
As reported by the Guardian I can’t see that a correction has been made for this (maybe it has - I’d expect any medical statistician to be extremely familiar with Bayes theorem).
I have to say that the ONS data has always struck me as somewhat inconsistent with the wider testing data, as I’ve previously commented.
My vote goes to you being put in charge of covid data collection and dissemination…with special dispensation to work from home in France.
9 Likes
You would expect the relevant corrections to have been made - the full study is reported to be in Clinical Epidemiology.
There is, also, another obvious explanation for inconsistency with wider test data, isn’t there? - if much of it is based on people that have symptoms, then if most people with covid don’t have symptoms, they have been testing mainly people with other viruses!
They mention it in passing but do not seem to properly account for it - perhaps because it blunts their figures.
The specificity of the test the ONS is using is possibly as high as 99.9% (this is discussed in the ONS publications) - so ~ 36 false positives in 36k subjects (they mention this, but do nothing with it).
If you subtract this from the 88 asymptomatic positives you get 52 - 27:52 or 65% asymptomatic is a lot of asymptomatic patients but it does not generate such eye-grabbing headlines.
Well, my concern was that if you went by the ONS data we were apparently picking up too high a percentage of cases, if you assume that there is a significant number of people who are infected but asymptomatic.
I should point out that worries about false positives in the ONS data has little to do with the mainstream testing, not least because the ONS survey was based on antibody testing IIRC.
PCR is typically quoted as specificity of 95% - admittedly false positives are less of a worry if there is a high pre-test probability of positivity (i.e symptomatic individuals) but it becomes more of a problem when increasing numbers of negative patients are getting tested - the worried well, contacts of other positive case who are, themselves, questionable.
The last figures were 14k cases out of 260k tests or about 5.45% - at that rate a 95% sensitive test is a problem because almost all of the positives could be meaningless.
The one thing that you can’t ignore, however, is the exponential rise in positive cases against, at best, linear increase in testing - so something is happening out there but I am not sure we are currently measuring it properly.
20k &10.4% de positiivé, aïe aïe aie.
Le taux de positivité des tests s’élève désormais à 10.4% en France
20.339 cas de coronavirus ont été diagnostiqués ces dernières 24 heures en France, rapporte Santé Publique France dans son dernier communiqué rendu public vendredi soir, un nouveau record. Par ailleurs, 62 personnes personnes sont mortes en 24 heures et le taux de positivité des tests s’élève désormais à 10.4%.
Wow! <27k quite a big jump from a couple of weeks ago! Stay safe.
Le nombre de nouveaux cas de coronavirus en France est sur une courbe ascendante. Mais cette fois, les nouvelles contaminations font un bond en avant conséquent : selon les données de Santé publique France publiées ce samedi, 26.896 nouveaux cas de Covid-19 ont été recensés en France en 24 heures. Un record depuis la rentrée. La veille, déjà, ce nombre avait franchi la barre des 20.000 cas pour la première fois. Le taux de positivité des tests augmente une nouvelle fois, pour s’établir à 11 %.
En 24 heures, 54 personnes sont décédées de la maladie. Ce qui porte le total, 32.684 décès.
While it is possible to argue (as I now would) that we are seeing something different to the situation back in March - either we are now overestimating rather than underestimating cases or, for whatever reason, we are seeing more mild cases, it is now inescapable that we’re on an exponential growth path again in the UK, France and much of Europe.
The UK is in a complete mess with local lockdowns, I’m not sure that they are the answer here (impossible to police, extremely hard to even know what the rules are in a given area, lack of public support given the antics of Cummings and his trip to Durham or Margaret Ferrier1) and I’m not sure they are being used any more effectively in France.
At least France did take a step back from uninterrupted rise, perhaps that brief period of success can be repeated but at the moment it just looks like the nation is playing catch-up.
1] I have a tiny bit of sympathy for Ferrier - when ill our natural instinct is to get home, where one feels safe (and I believe that she was in London alone) - but she should now resign otherwise how can she demand anyone else abides by the rules when she did not.
1 Like
Yes the instinct is to get home if you are ill, but she shouldn’t have left home in the first place having had a Covid test. I got home from mine and ordered a charger cable from Amazon, rather than getting one much cheaper from the shop across the road
Ah, I must admit I thought she’d had the test and the result in London - OK, small amount of sympathy for wanting to return home but zero for travelling when symptomatic & waiting for a result and doubly less than zero for then, apparently, going to church and giving a reading after the positive test (which I just saw checking the original timeline).
On that basis I’m afraid she really deserves to be suspended from Parliament, recalled and deselected.
2 Likes
did I hear that Doris was calling for her to be called to account… pot kettle Cummings?
A bit worrying…
But how much virus were these surfaces blasted with, and were they kept in perfect conditions? Previous studies have been shown to be unrealistic as liberally doused surfaces with far more virus than would be deposited by a few sneezes… and if in the dark then no chance of it being killed by UV light.
2 Likes
As Jane says - for every expert opinion there is an equal and opposite expert opinion…
https://www.thelancet.com/pdfs/journals/laninf/PIIS1473-3099(20)30561-2.pdf
2 Likes
The article itself is OK - points out that the 28 days survival was rare, required darkness, constant temperature, etc, and that surfaces play a very small role in spreading anyway - but this just begs the question of why the alarmist headline? (I know the usual explanations - clickbait, etc - but should the BBC be indulging in this when many people are already very worried?)
1 Like
Hmmm, a record breaking day - just not in a good way.
Not that the UK is doing any better mind you 
1 Like
That’s our toubib’s view as well. I’m still going to get vaccinated. Not that it did me much good last year. My daughter reports that life continues as usual in Perth. At least I don’t need to worry (too much) about her.
I’ve got one booked at 07:50 tomorrow - slightly worryingly several people have said this year’s has made them feel unwell for the first time - I guess we’ll see how it goes.
I had mine on Tuesday, no ill effects so far.