FYI, the latest on Test, Trace, Isolate from the Scottish Government
https://www.gov.scot/publications/coronavirus-covid-19-test-trace-isolate-support/
thats an appropriate use of "contacts of confirmed cases must self isolate".
doing it for all random passing contact between strangers is going to lead to most of the country quarantining.
From the Scottish Gov link:
9. For the purposes of contact tracing, a close contact is someone who has been physically close enough to the confirmed case for a long enough period of time, that they may have had the infection transmitted to them. For COVID-19, this includes everyone who has been less than 2 metres away from a confirmed case for 15 minutes or more. The risk of the disease being transmitted is higher the closer the contact, the greater the exposure to respiratory droplets (for example from coughing), or the longer the duration of the contact.
15 minutes seems pretty long? It'll obviously trigger if you've been sat next to someone all day at work, but if you happen to be in the supermarket with them at the same time?
it needs the speed of testing to be quicker than the speed of incubation to transmission.
Yup, and that's never going to happen. It would require everyone who has a temperature or cough to get seen, diagnosed and tested the same day. No system can do that. 50pc of the population would need to be medical staff diagnosing the other 50pc! (I exaggerate, but you get the point.)
No silver bullet, just load of poo bullets that together might keep R below 1.
What you NEED is a model with descriptive statistics such as population density, oecd spending on health per capita, age breakdown in the population, smoking status, cardiovascular health by age, number of patients in nursing homes, days since lockdown, factors of lockdown, adherence measures… Whilst association is not causation, one can make some pretty informed guesses about relevant covariates. That’s the fun bit.
I see if I still have a 20 year old version of Canocco. I'm sure a limited understanding of how to analyse plant community data from this time would be sufficient....just need the data sending my way! Bit of dca and some monte carlo permutations and bob is your mothers brother.
i think i used the wrong words. Ignore the first person coughing for now (you would need a time machine for that), think of the next person down the line. whats the time between them being coughed on, and the virus taking sufficient hold to replicate enough, and then start shedding?
It needs (imo) to be only confirmed cases get to flag up on the app; and their contacts get given tests (but only quarantine with symptoms or positive test).
I agree.
Both Madame and I had symptoms that could have been CV19 in early part of France's lockdown, there was no point telling anyone apart from our GP when Madame's conjonctivitis was so bad she had trouble opening her eyes and needed some eye drops (Whatsapp consultation with GP, prescription dropped out of GP's window, pharacist dropped meds in Madame's bag). We were in confinement so no contact so no threat to anyone.
Would we have signed up to an app that offered an immediate no questions asked test and positive result before being added to the database? yes. Would we have signed up to an app under any other conditions? no. The app needs to be a positive for the user with no possible downsides.
As for tracing contacts, there's no point putting them on the data base unless they test positive.
I think it's useful for everybody to know who has been tested positive recently, I think it would be pointless adding me to the database because my seasonal alergic asthma cough has been worse than usual - but that's what you'll get with an app that flags people with symptoms that haven't been tested or contacts that haven't been tested. You won't be able to move without your app flagging people up unless it's test based.
Edit: see the "have you had it thread". you can count the positve tests in the thread on the fingers of one hand. The rest really don't know. I haven't even posted on the thread because I simply don't know and the symptoms could be a host of other ills. A blood test for antibodies would be useful in Madame's case and some of the people on that thread who could make decsions based on the result that would be economically useful to the country/economy. Madame would volonteer for high risk work if she knew she had antibodies.
I don’t understand why UK Gov haven’t gone for the decentralised (Apple Google) approach.
100%.
Just to be clear, Google/Apple arent working on an app as such, its just a framework of APIs that allow app developers to get easy interoperability and functionality. This framework works on a decentralized approach, similar to this https://github.com/DP-3T/documents .
The Google/Apple/Dp-3T solution is pretty smart, and does give good privacy with no real identifiable info stored or sent to the health authority.
Which is why the UK decision to not use it is so fishy. They not only now have to convince people that their data is safe with private companies they also have to work around any issues with hardware/software that the Goog/Appl solution solves from the outset.
Germany gave up: https://uk.reuters.com/article/uk-health-coronavirus-europe-tech/germany-flips-to-apple-google-approach-on-smartphone-contact-tracing-idUKKCN22807X
And Australia coudln't get it to work on iPhones without the app being open:
https://twitter.com/MikeHaydon/status/1254646784825061377
The Govt had a privacy wise decent solution supported by the vendors ready to go, and yet they chose the harder path of doing their own. You really need to ask yourselves why that is.
if you happen to be in the supermarket with them at the same time?
Nothing new here really. The risk of infection from being within 2 metres of someone for a few seconds/couple of minutes is, and always was, vanishingly small.
Good to know that the likes of Joanna Cherry is looking at the data privacy implications of the UK Govt. app.
from the bbc article in the EU travel thread
Some have gone further, though, with Emirates offering passengers rapid Covid-19 blood tests prior to boarding at Dubai airport terminals. Emirates says the tests produce results within 10 minutes.
presumably there's a big machine with all the fancy chemical and electrical shizzle sitting there that does the test - rather than a litmus paper/pregnancy test type thing; but still game changing compared to postal tests.
The Govt had a privacy wise decent solution supported by the vendors ready to go, and yet they chose the harder path of doing their own. You really need to ask yourselves why that is.
Money and power to Dom and pals. You don't need to be Sherlock Holmes 🙂
There's no shortage of tests but some really aren't worth using given their lousy reliability, 278 live virus tests, 233 antibody tests of which 84 are quick tests have been approved in the world:
Au 22 avril 2020, 278 tests de RT-PCR visant à déceler la présence du SARS-CoV-2 dans des sécrétions nasopharyngées et 233 tests sérologiques permettant de révéler celle d'anticorps dans le sang, dont 84 tests rapides, ont été commercialisés dans le monde
SAGE members just been published, although with a caveat...
'These meetings are also regularly attended by officials from Her Majesty’s Government. These attendees have not been named.
Permission to publish names was requested from all participants. Those who did not give permission have not been named.'
Big flaw in the ScotGov strategy - they're expecting anyone notified through the process to self-isolate for 14 days. I suspect that anyone who thinks they've already had, and fought, the virus will just ignore this.
Plus there would likely repeats of the 14 days as you come into contact with various folks. Possibly more acceptable than a permanent lockdown but difficult to plan for as both an employee and an employer.
plant community data from this time would be sufficient
You do know that the foundations of modern statistics were laid down at Rothamsted by Ronald Fisher et al? You'd be in very good company 😉
https://en.wikipedia.org/wiki/Rothamsted_Research#Statisticians
a year of lockdown wont happen.
the untimely death of 100000’s, and a year of misery, unemployment and poverty for many millions due to restrictions on certain things will be the result if the variety of measures don’t work.
there is a point (which I hope is not close, and never occurs) where a quarter million deaths is the lesser of two evils.
when your time frame is a year, pretty much every job apart from the leisure industry becomes “essential” and the daily life of people cannot be paused for that long without drastic consequences.
Maybe I meant an adjusted lockdown strategy. Keep tweaking the measures until it's sustainable. Otherwise, the worry is we're into 1918 second-wave territory.
SAGE members just been published, although with a caveat…
Interesting list. Can't see Home Office or MoJ on there. Bright lad deep mind Demis you'd think would know a bit about apps. Can't see primary care or social care expertise.
Not naming officials in attendance is usual. These are often relatively junior folks. Senior officials who are members of the group are named.
I think you only have to self-isolate until your test result comes back - for a maximum of 14 days if you have no symptoms/negative test in that time - but I could be reading that wrong....
Those who do test positive for <abbr title="Coronavirus Disease 2019">COVID-19</abbr> will be asked to isolate for 7 days and their close contacts will be asked to self-isolate for 14 days. It will be important that the people identified as close contacts remain in self-isolation for the full 14 days
Testing is still only for the symptomatic though?
As I say, you could end up with multiple consecutive 14 day isolation periods.
But what does STW think of phone based apps for ‘contact tracing’?
Nope, won't be installing any such thing
Surely only if we accept that you can catch it more than once?
You have a close contact - so you have a test - it comes back negative and you are free to roam the streets again. If this takes 14 days then that's pretty hopeless.
You have a close contact - so you have a test - it comes back positive. Chances are you recover and that's you free to roam the streets again forever.
(once we know that you can't catch it twice - I am sure I read somewhere that the people who have had it twice are now thought to be false positives)
I've found the Independent Sage thing:
(I'm 23 minutes in so far and it's not as good as I hoped.)
Okay, I owe a huge apology to whoever I was rude to when they talked about the possibility of this a while back. It sounded like conspiracy theory nonsense back then. Again. Sorry.
When the NHS app uses the Apple/Google framework, I’ll use it. No way in hell this lot are installing a back door on any of my computers, handheld or not.
@hels - or, you have a close contact, they test positive, you isolate for 14 days but haven't been infected. Day 16, another close contact is flagged and you're back in the house for another 14 days, etc etc etc.
When the NHS app uses the Apple/Google framework, I’ll use it. No way in hell this lot are installing a back door on any of my computers, handheld or not.
Not sure I trust Apple and google either though! Oh dear!
When the NHS app uses the Apple/Google framework, I’ll use it.
My understanding is the Apple/Google bit optimises the Bluetooth LE (so it can run reliably in the background) and provides the contact-matching, key sharing and notification bits. Your health service can write what it wants on top of it, so don't just assume if it's using the Google/Apple APIs that it's legit.
don’t just assume if it’s using the Google/Apple APIs that it’s legit.
100% correct. Looks like they (Google/Apple) are imposing some T&Cs around the apps that use the API though - however, we know that our current Government's inner circle wipe their arses on the rules.
COVID19-STATIC.CDN-APPLE.COM Lawyers initial opinion is that the NHS app data sharing will be unlawful.
https://twitter.com/RaviNa1k/status/1257212168191324160
Not sure I trust Apple and google either though! Oh dear!
This matters less with the decentralised model (unless you think they are lying about it being decentralised or that it leaks more info than stated)
Jamze described it nicely above:
Imagine everyone was carrying around a bag of raffle tickets. Every time you meet someone (for a defined time/distance) you both secretly swap a ticket and keep the stub.
When somebody tests positive, that person’s raffle ticket numbers are broadcast out and anybody with a match is flagged as a contact. You have no idea who the person was who tested positive, or who else has been contacted.
Your raffle ticket number randomly regenerates every x amount of hours.
A more indepth doc explaining it is here:
The centralised model only makes sense if the health authority really do want more info (location,time,time in contact,phone number.... whatever else) which might be valid depending on your point of view.
This is backed up From the BBC article
But NHSX believes a centralised system will give it more insight into Covid-19's spread, and therefore how to evolve the app accordingly.
"One of the advantages is that it's easier to audit the system and adapt it more quickly as scientific evidence accumulates," Prof Christophe Fraser, one of the epidemiologists advising NHSX, told the BBC.
DP-3T is open source, so would be the preferable solution.
Plenty of other people concerned about the centalised model and its use of private companies to handle the data:
https://www.theregister.co.uk/2020/04/29/academics_open_letter_nhs_coronavirus_app/
Not sure I'm willing to sacrifice a potentially important tool in getting on top of the virus spread to enable lockdown to be eased on the altar of data privacy.
My issue is how much time and public money the government will waste on something that won't work at the end of the day.
I really don't give a damn who sees what is on my phone if it gets us closer to beating the virus. Google/Facebook etc have had my data for years, and apart from some slightly be left field adverts from Wish, I seem to have survived the experience.
Not a cat in hells chance that i or anyone in our business will be downloading that "Cummins Driven" data scraper...
Social engineering on a monumental scale... wonder what the DPIA looked like? (Data Privacy Impact Assessment) assuming there is one?
National interest my a**e
Sorry morecashthandash but that "attitude" is why we have brexit just over the hill from Covid 19.
Do you have any idea what the commercial and political value of that data set is? Probably not a big enough number available.
National interest my a**e
So if one of the 18,000 contact tracers called you up and said 'oldmanmtb2 it's been identified you've been in contact with someone who has tested positive' would you refuse to take that call? They're going to want similar info from you.
I really don’t give a damn who sees what is on my phone
Ok. What about your medical data being made available to a 3rd party with no oversight? Under cover of developing a “Covid app”.
something that won’t work at the end of the day
It won’t work if we don’t use it, and we won’t use it if we can’t trust it.
Social engineering on a monumental scale… wonder what the DPIA looked like? (Data Privacy Impact Assessment) assuming there is one?
Something along the lines of immoral but profitable?
Something along the lines of immoral but profitable?
First one is always free.
https://tech.newstatesman.com/coronavirus/palantir-45-engineers-to-nhs-covid-19-datastore
They wont be calling me Jamze as i won't have the app...
If your happy to use it crack on.
They wont be calling me Jamze as i won’t have the app…
But they will if we are manually contact tracing in parallel. I assume this is the plan, the app is just one element.
Well from the answer to the privacy question just now on the briefing, they've described the distributed design (same as Apple/Google). They also said you order a test from the app once you get symptoms.
So for the non installers, what would you need to know to make you comfortable to install it? Is it the concept of (anonymised) contact tracking, the fact the government are involved, that Cummings was somehow involved somewhere?
How will "they" know who oldmanmtb2 is or what his phone number is if he hasn't got the app?
How will “they” know who oldmanmtb2 is or what his phone number is if he hasn’t got the app?
They interview the person who tested positive.
How will “they” know who oldmanmtb2 is or what his phone number is if he hasn’t got the app?
The same way they knew when there was contact tracing at the early stages of the outbreak. His friend tests positive, and then provides details of the people he's been in close contact with over the last week. Of course he could refuse to give that info too.
Okay, I owe a huge apology to whoever I was rude to when they talked about the possibility of this a while back. It sounded like conspiracy theory nonsense back then. Again. Sorry.
Apology accepted.
Is it the concept of (anonymised) contact tracking
No. And if that was all that was to happen, then it wouldn’t work anyway. The data stored needs to be able to return who (or at least which device) was involved for this to work. I accept that the data will be used to identify and contact people, but, because of that necessary lack of true anonymisation, I want to know and be in control of what data is taken from me and when. A closed app that can be used for real-time and evolving data harvesting beyond a specific set of parameters, stored centrally with no promise of use being restricted to the NHS and other public health bodies. No way.
So for the non installers, what would you need to know to make you comfortable to install it?
That no Tory politician, or shady behind the scenes unaccountable 'adviser' had been anywhere near it. I know we've been ignoring party politics for the duration, but Tories... These people have been messing up my life for pretty much all my life, and I see no reason to think that they have my best interests at heart now.
