NHS contact tracing app

[bert]

There's a manual parallel system planned. Don't know any details.

[bert]

It doesn't track you. It doesn't track you. It doesn't track you.

It doesn't track you.

[Joshua Snow]

It doesn?t work like that. It works on someone testing positive to the virus, not just saying they don?t feel great.

[Dave Liquorice]

Fraction of second if an infectious person coughed into your face...

In the real world if just walking past (not touching) someone got you infected the R0 wouldn't be between two and three, it'd far, far higher. And the rate of transmission so fast we'd still be trying to work out WTF was going on with a majority of the global population infected.

The virus needs to be physicaly transported person to person. Be that via shared surfaces (handled and put back goods, trollys/baskets, self-checkout touch screens, chip 'n pin pads, door handles, hand rails, lift buttons etc) or droplets from someone's cough/sneeze, probably onto a surface that is subsequently touched rather than breathed in. Which is why wash/sanitise your hands frequently (when out and immediatly on return home) and don't touch your face are so important.

One reason I don't like the one way systems some supermarkets have put in place. Some one dithers, next person doesn't want to pass, so stands in 2m proximity of that person, the next perhaps does the same. Before long you have a little knot of people all waiting for some one to cough... Far better to have free flowing movement of people even if they break the 2 m rule for a few seconds. IMHO you want, on average, to minimise time close to and maximise distance from others.

Not in epidemilogical terms. Packed into public transport, all day in an office, crowded pub or restraunt for the evening, sports crowd, theatre/cinema audiance where you are in close proximity of people for an extended period of time are the situations when transmission is most likely to occur. Not just walking past some one on the street.

Remember this app is really about gathering information about how fast (and loosely where) the virus is spreading. It's not about accurately telling people they have been infected, it can't do that, though it might get quite good as the modelling evolves over time. The user alerts aren't required, that's just the carrot to get people to install it.

One would assume so as they are using the signal strength, measured at 8 second intervals, as one of the metrics to guage the risk of transmission for a given proximity event. The registration also involves telling the server what the device is to compensate for differnt radio sensitivities.

Up and past so quick I doubt it'll be seen.

Not relavant to the NHS app. It doesn't have any location info other than a partial home postcode (optionally?) entered at registration.

The app is a broad epidemilogical tool, so "they" can see the rate of transmission. Ease some of the restrictions, watch the rate (almost in real time), if it starts to look as if it's begining to take off "they" can reimpose rescrictions or delay further easing.

[Dave Liquorice]

devices at

Not the NHS app. Initial user "I'm ill" report is based on self diagnosis by answering questions put by the app. The google and/or apple systems seem to rely on a positive test.

Read the .pdf about the NHS app recently linked to.

[Joshua Snow]

The bluetooth distance function doesn?t work by signal strength.

[Clive Page]

And Bruce Schneier doesn't think any of these apps will really work in practice:

[Clive Page]

The technical note does deal with this and other problems e.g. a school pupil could try to get everyone in their class locked down and hence get another couple of weeks off school. They have some ideas on how to avoid this but I didn't find them all that convincing.

For details, see

And let us know here whether you are convinced or not.

[alan_m]

The SIM is allocated a phone number and the phone has a digital identity. Most people either have a contract or have registered their phones with a mobile supplier so why do they need a postcode at all. They just get your address from the information provided by the mobile phone supplier.

[Clive Page]

I agree with that. But I'm not too worried about what happens in these passing encounters in a supermarket because of the following:

It seems to be generally accepted that the average value of R0 is 3 or 4. The average person who gets infected may well be infectious for several days or even a week, so they probably come quite close to a whole lot of others in that time. Despite that, on average then only manage to infect 3 or 4 others. That must mean that the risk for a single fairly brief encounter isn't that high, unless someone else actually sneezes or coughs directly at you. And probably 90% of supermarket customers are not infected, at least not yet.

But we still don't know what the risk is from touching supermarket trolleys, or other hard surfaces. That may be a more serious problem, who knows. Unfortunately until each trolley is fitted with a Bluetooth beacon - we won't know.

[Clive Page]

Can't. It is somehow amusing that the Dartford Crossing charge service is not even in Beta testing yet - still says Alpha and it must be three years in public service at least.

[Joshua Snow]

Just allows the app to be installed without the user being asked if it should be allowed to use the bluetooth.

No data goes to apple or google.

>

[Dave Liquorice]

registration.

GDPR won't (shouldn't!) allow the mobile operators to release that information for this reason, without express consent.

Anyway the system doesn't need location information, it's just a "nice to have". So "they" can look at the numbers of new cases each day in each postcode district and spot potential hotspots and thus get ready for a possible localised upsurge in hospital admissions in around 10 days.

[Dave Liquorice]

Don't like that idea at all. Thin end of the wedge... If an app, any app, wants acess to anything it needs the user to grant the permission. If they refuse it's up to the developers to decide how to handle that. Explain why the acces is needed, explain that the app won't work or have limited function without, etc.

Something must get sent to apple/google otherwise how do other devices get told or be given information to work out and raise a notification?

[Joshua Snow]

Its more complicated than that with a contact tracing app which needs to ensure that the worst of the technoklutzs don't stuff up the installation so that it can't work.

The anonymised id is the only thing that apple or google sees.

Only the NHS can turn that into anything useful.

[Dave Liquorice]

And that's *without* social distancing.

And they are infected and contagious.

And will probably stay like that. There doesn't seem much about how the 80% of the population who get infected and only have "mild symptoms". I do know that a temperature of 38+ C is going to put you in bed or at least semi-comatose on the sofa. Going shopping is not likely to be high up the priority list. Shortness of breath likewise,

The big unknown is how long (if at all) people are contagious before they recognise the symptoms and self-isolate assuming they show any symptoms at all.

Supermarkets I go to minimal, handles of trollys and baskets are cleaned after every customer.

That's more of a problem but noting when your hands could be "dirty" and then not touching your face and remembering what you have touched with those "dirty" hands so when you retouch those objects with "clean" hands you set the "dirty hands" bit. Your wallet/purse, payment cards, phone, keys etc are the objects that'll trap you rather than the "obviously dirty" shopping.

When using the car I only use the drivers door or touch any controls after cleaning my hands, to keep them "clean". I clean my hands everytime I return to the car. The cleaning stuff I access via the "dirty" passenger door but thinking about it as I type. Why am I making the passenger door "dirty" when I've just made the boot "dirty" putting the shopping into it? Cleaning stuff should be in the boot... but I can't close the boot with my elbow. It does have two pull handles but will I reliably remember which is "clean" and which is "dirty" and use the appropiate one for my hand state?

[John Rumm]

No, indeed. (its debatable even if bluetooth is not going to be too broad a brush)

I think in built up areas with lots of base stations, the resolution from triangulation is reasonably accurate - say tens of meters. Certainly good enough to profile your daily routine, where you work, shop, visit etc.

and sainsbury's probably know what you are going to buy before you even get there!

[Andy Burns]

Using ESSI information, presumably the BT/BLE APIs provide this?

The NHS app also notes the device model at time of first run, so that they can scale up or down the estimate of range from RSSI, based on known performance of different types of radio.

[Joshua Snow]

But some might well do before its that bad.

That is known and the fact that some never get any symptoms and have infected lots of people.

[Andy Burns]

I was talking about the new BLE tracking features that Apple and Google are adding to the base O/S to allow covid apps to do things that normal apps aren't allowed to do ... I daresay that "footfall trackers" would love to find a backdoor way of accessing that information, and tying it to your search engine/facething account.