Would you install and use an NHS Covid tracking app?
Poll: Would you install and use an NHS Covid tracking app?
Total Members Polled: 875
Discussion
pip t said:
Zirconia said:
Thanks. One thing that stands out to me from that article is:Article said:
NHSX was not aware of this project beforehand, but now plans to integrate the technology into its own product.
The fact they'd already started creating the app and are now integrating the Apple/Google system doesn't fill me with confidence in it being implemented entirely as published by the tech companies. Plus the fact that the NHS may use data for 'care and research' - if it's implemented properly it doesn't generate that kind of data....However the fact they're committed to open sourcing it does guard against there being anything sneaky in it I guess....
The yellow/ red alert system seems like it has potential in terms of differentiating between self diagnosed and confirmed cases, but leads to the question of what action are you supposed to take on receiving a 'yellow' alert? Are you still supposed to go down the full self isolation route? Because that would leave it wide open to abuse if you can self declare and cause others to have to quarantine on the basis of it...
But usage wise, does seem to be a bit hit and miss and relying on user being honest which will also lead to other doing what they will to spoof it, lie etc. Problem then is how do you make it honest and foolproof. Get a red and I expect a lot will panic a bit, including me.
I expect more details might provide a better insight.
And yeah, clocked the NHSX bit.
steveo3002 said:
i wouldn't install
but then my phone often thinks i'm somewhere i'm not , and apps crash the time so good luck with it working well
If it works as I understand it does, it doesn't need to know where you are, or indeed care where you are.but then my phone often thinks i'm somewhere i'm not , and apps crash the time so good luck with it working well
You could be in the street, in the pub, on the moon. Your phone transmits an encrypted key, unique to you, but containing no information about who you are. As do other phones in the same area, if somebody who is in that area later tests positive for the virus, they can then track all the people they have been in contact with using that list of unique keys.
Zirconia said:
Few problems, open source on the app is not the sole issue, it is the integration with other patient data and how that is all used or abused. That needs to have the process fully open to scrutiny including any tendering processes. Meaning not just the app needs to be open, what else are they doing.
This is the thing though, and why I keep banging on with the phrase 'if implimented as designed.' If it's implimented as designed, it can't be integrated with any other patient data, simply because it has no idea who any individual user is. That's where the open sourcing is vital, so that it is open to auditing to check that it isn't doing anything sneaky in the background that could identify users.I completely take your point that there's all kinds of fishy stuff that could be and may be going on with patient data sharing etc, but if done properly, this particular system could never be a part of that.
Megaflow said:
If it works as I understand it does, it doesn't need to know where you are, or indeed care where you are.
You could be in the street, in the pub, on the moon. Your phone transmits an encrypted key, unique to you, but containing no information about who you are. As do other phones in the same area, if somebody who is in that area later tests positive for the virus, they can then track all the people they have been in contact with using that list of unique keys.
Indeed. And I think part of the 'marketing' for this when it is finally released has to NOT use the word tracking. Quite understandably, a lot of people will misinterpret that as location tracking, which, again understandably, would dissuade a lot of people from using it.You could be in the street, in the pub, on the moon. Your phone transmits an encrypted key, unique to you, but containing no information about who you are. As do other phones in the same area, if somebody who is in that area later tests positive for the virus, they can then track all the people they have been in contact with using that list of unique keys.
One thing this app emphatically does not do is track anyones location.
Not worried about the app in the slightest (as described it is pretty benign) .... but;
- Bluetooth is the most random range finder ever invented. Walking 50 yards from someone in clear air will generate the same signal as sitting next to someone with a phone in their pocket. One has infection potential, one doesn’t.
- It assumes that everyone has their phone with them at all times - maybe the yoof do, but a lot of people don’t.
- It’s going to have to be really clever to work out the difference between two people passing on a pavement and two people sitting in a traffic jam.
IMO as soon as it goes live, everyone will be warned they have been close to an infected person.
- Bluetooth is the most random range finder ever invented. Walking 50 yards from someone in clear air will generate the same signal as sitting next to someone with a phone in their pocket. One has infection potential, one doesn’t.
- It assumes that everyone has their phone with them at all times - maybe the yoof do, but a lot of people don’t.
- It’s going to have to be really clever to work out the difference between two people passing on a pavement and two people sitting in a traffic jam.
IMO as soon as it goes live, everyone will be warned they have been close to an infected person.
Quite an informative article on it from Wired here: https://www.wired.com/story/apple-google-bluetooth...
rxe said:
Not worried about the app in the slightest (as described it is pretty benign) .... but;
- Bluetooth is the most random range finder ever invented. Walking 50 yards from someone in clear air will generate the same signal as sitting next to someone with a phone in their pocket. One has infection potential, one doesn’t.
True to an extent. Signal strength is a vague way of identifying range, but yes, not all that accurate. Bluetooth 5 has the ability to work in two range 'modes' - longer range but slower data transfer, or shorter range with higher data transfer - but again, this isn't hugely accurate, and is only available on recent model phones.- Bluetooth is the most random range finder ever invented. Walking 50 yards from someone in clear air will generate the same signal as sitting next to someone with a phone in their pocket. One has infection potential, one doesn’t.
rxe said:
- It assumes that everyone has their phone with them at all times - maybe the yoof do, but a lot of people don’t.
- It’s going to have to be really clever to work out the difference between two people passing on a pavement and two people sitting in a traffic jam.
There is also a duration aspect - you have to be within a phones range for 10/15 minutes for it to log a contact - so just passing someone in a street wouldn't trigger it. This also to an extent helps with your first point.- It’s going to have to be really clever to work out the difference between two people passing on a pavement and two people sitting in a traffic jam.
rxe said:
IMO as soon as it goes live, everyone will be warned they have been close to an infected person.
That won't happen - when it first goes live there won't be anything in server to cause the triggers. As it build up though, yes, there's much scope for false positives.It's very very far from perfect, you're right. There's all kinds of issues to work through. It is however a method of doing it without going down the road of monitoring peoples geographic location, which makes the whole thing far more palatable to the mass of population who would need to adopt it for it to have any effect.
pip t said:
That won't happen - when it first goes live there won't be anything in server to cause the triggers. As it build up though, yes, there's much scope for false positives.
It's very very far from perfect, you're right. There's all kinds of issues to work through. It is however a method of doing it without going down the road of monitoring peoples geographic location, which makes the whole thing far more palatable to the mass of population who would need to adopt it for it to have any effect.
The duration thing avoids an avalanche of false positives, but surely renders the whole thing pointless? If I get in a lift with someone coughing their guts out for 15 seconds, that’s very high risk, but according to this app, not a contact. Ditto someone coughing on the street next to me. It would be good for “you’ve been stuck in a meeting with a sick person”, but for the majority of the population, that’s a poor use case.It's very very far from perfect, you're right. There's all kinds of issues to work through. It is however a method of doing it without going down the road of monitoring peoples geographic location, which makes the whole thing far more palatable to the mass of population who would need to adopt it for it to have any effect.
Downward said:
Pointless unless 100% say yes surely ?
Not really. In the same way as the lockdown measures only need 75-80% compliance to work, this would only need the majority to have to it have an effect. Sure, if 100% had it it would have a greater effect, but that's impossible given not everyone has a compatible phone, not everyone carries their phone everywhere etc. I believe one article gave a figure of 60% needed to make it work effectively. Granted that's still a metric s
tload of people who need to use it though!rxe said:
The duration thing avoids an avalanche of false positives, but surely renders the whole thing pointless? If I get in a lift with someone coughing their guts out for 15 seconds, that’s very high risk, but according to this app, not a contact. Ditto someone coughing on the street next to me. It would be good for “you’ve been stuck in a meeting with a sick person”, but for the majority of the population, that’s a poor use case.
Undoubtably. But hopefully the person coughing their guts up in a lift would be aware they had symptoms and be self isolating, meaning they shouldn't be in the lift in the first place.Where it really could work is in asymptomatic/pre-symptomatic transmission. For example you go for coffee with someone, your phone logs that you've had contact with them. 3 days later, they start experiencing symptoms, and they log it with their phone. Because you've had contact with them in the pre-symptomatic phase, your phone then alerts you to something you wouldn't have been aware of.
Edited by pip t on Monday 13th April 14:14
Edited by pip t on Monday 13th April 14:15
Downward said:
Pointless unless 100% say yes surely ?
No.I'm not a mathematician but I'd assume there's a curve or chart or something against people participating against impact.
No idea where the sweetspot is but I wouldn't expect it's 100% as the effect of it won't be all or nothing.
pip t said:
Undoubtably. But hopefully the person coughing their guts up in a lift would be aware they had symptoms and be self isolating, meaning they shouldn't be in the lift in the first place.
Where it really could work is in asymptomatic/pre-symptomatic transmission. For example you go for coffee with someone, your phone logs that you've had contact with them. 3 days later, they start experiencing symptoms, and they log it with their phone. Because you've had contact with them in the pre-symptomatic phase, your phone then alerts you to something you wouldn't have been aware of.
But why 10 minutes? If I got in the lift with that person, I could have just as easily caught it. If I (say) go into London to meet someone for coffee, I probably have glancing contact with 1000 people. This app will tell me if that one person is ill .... not the cast majority of people I walk past/near/sit next to etc.Where it really could work is in asymptomatic/pre-symptomatic transmission. For example you go for coffee with someone, your phone logs that you've had contact with them. 3 days later, they start experiencing symptoms, and they log it with their phone. Because you've had contact with them in the pre-symptomatic phase, your phone then alerts you to something you wouldn't have been aware of.
Edited by pip t on Monday 13th April 14:14
Edited by pip t on Monday 13th April 14:15
rxe said:
But why 10 minutes? If I got in the lift with that person, I could have just as easily caught it. If I (say) go into London to meet someone for coffee, I probably have glancing contact with 1000 people. This app will tell me if that one person is ill .... not the cast majority of people I walk past/near/sit next to etc.
I guess there's an element of "pick a number" and an app can't tell things like if you're in an enclosed space.Hell you could meet someone for 10 seconds but if they sneeze in your face it's not good.
pip t said:
Completely aside from the topic at hand, and obviously I completely respect your decision not to own a smart phone, but there are some extremely clever ways they can help people with hearing loss. Google particularly has done great work in creating a system that can provide captioning for user generated content (Eg if your family send you a video, your phone can provide captioning for it), and they're working on applying the same technology to live phone/video calling. That may be of interest? There's also ways that both Apple and Android phones can directly connect to hearing aids (Obviously depending on your hearing aid type). As I say, clearly if you're not interested in owning a smart phone that's entirely your prerogative, but there are ways they can help hearing loss.
Thanks for your interest Pip, and I appreciate this advice.The thing is, though, I am retired and simply don't get out that much nowadays. A bout with cancer a few years back has left me appreciating a simple, home based life. Having a continuously high level of connectivity to the Internet is of no interest to me. I have a PC and an iPad at home and use them a lot. Nothing ever comes along that means I would need facilities like this on the occasions that I am out. I can appreciate that people who are younger, still working and get about more would find all that a smartphone can offer is invaluable, and I can see that they are enormously clever devices, but they are not for me.
It's like I said previously, there's just no gap in my life that a smartphone would fill.
A couple of birthdays ago I mentioned to my family that I am starting to get a little concerned on the few occasions I am out driving on my own that I might need a phone if I should get in trouble. My son bought me a simple phone, on a PAYG card that does little more than make a call. I think I have turned it on three times since, just to call my home number and see if it works.
One or two posts on this thread, from I assume younger contributors, seem to indicate that they find it strange that other people might have different circumstances and needs from themselves. Personally, I find the sight of people wandering about with their phones poised in front of their faces all the time equally odd, but we all live our lives the way we want to.
rxe said:
But why 10 minutes? If I got in the lift with that person, I could have just as easily caught it. If I (say) go into London to meet someone for coffee, I probably have glancing contact with 1000 people. This app will tell me if that one person is ill .... not the cast majority of people I walk past/near/sit next to etc.
I think the 10 minutes came from a study showing the likelihood of transmission. If you just have a 'glancing blow' with an infected person the likelihood of transmission remains relatively low. A sustained amount of time dramatically increased the likelihood of transmission.In the end its all a balancing act. Yes, logging every single contact regardless of time would increase the amount of transmission points notified. It would also increase false positives, etc etc. Requiring a set duration reduces the likelihood of false alarms, but, yes, also reduces the potential real alarms. You have to have a balance somewhere to keep the system practical and effective.
It's another case of don't the perfection be the enemy of the good.
I would use it but I think it could be better implemented.
It should assign a probability score - this would increase depending on how close you are to someone, their probability score and how long you are close to them and decay over time.
If it's greater than a "red" threshold, then the course of action you should take should be made clear - get tested (if tests are available) which would reset your score to 0 or 1, or if no tests are available self-isolate (with support available so you or your employer do not lose out and you are not tempted to hide the fact you are at risk).
If it's greater than an "amber" threshold, then you should do what you can do at the moment (e.g. work but work from home if you can, go shopping for essentials only). You should not work however if your work involves caring for the vulnerable. You would not be able to go out with friends etc.
If it's less than this, then you can behave relatively normally, go out with friends etc. You might be asked to show a phone with a green score if you want to go into a theatre etc.
It should also report an approximate region you are in - if the average score in a region is higher than a threshold, that region could be locked down but other regions of the country kept open.
For people without smart phones, there should be an option to be provided with a device free of charge (configured so it can only be used for this purpose).
It should assign a probability score - this would increase depending on how close you are to someone, their probability score and how long you are close to them and decay over time.
If it's greater than a "red" threshold, then the course of action you should take should be made clear - get tested (if tests are available) which would reset your score to 0 or 1, or if no tests are available self-isolate (with support available so you or your employer do not lose out and you are not tempted to hide the fact you are at risk).
If it's greater than an "amber" threshold, then you should do what you can do at the moment (e.g. work but work from home if you can, go shopping for essentials only). You should not work however if your work involves caring for the vulnerable. You would not be able to go out with friends etc.
If it's less than this, then you can behave relatively normally, go out with friends etc. You might be asked to show a phone with a green score if you want to go into a theatre etc.
It should also report an approximate region you are in - if the average score in a region is higher than a threshold, that region could be locked down but other regions of the country kept open.
For people without smart phones, there should be an option to be provided with a device free of charge (configured so it can only be used for this purpose).
pip t said:
The fact they'd already started creating the app and are now integrating the Apple/Google system doesn't fill me with confidence in it being implemented entirely as published by the tech companies. Plus the fact that the NHS may use data for 'care and research' - if it's implemented properly it doesn't generate that kind of data....
They can build some optional stuff in to help them analyse spread of the virus without compromising patient confidentiality.For example, when you set up the app, it could ask you to enter: age, sex, ethnicity, first three letters of home postcode, first there letters of work postcode. Those could all be optional but would help track demographics and hotspots in the virus. It could also be only released to the NHS if you test positive. Lots of stuff could be built in without identifying individuals or locations that is helpful in the fight.
AstonZagato said:
They can build some optional stuff in to help them analyse spread of the virus without compromising patient confidentiality.
For example, when you set up the app, it could ask you to enter: age, sex, ethnicity, first three letters of home postcode, first there letters of work postcode. Those could all be optional but would help track demographics and hotspots in the virus. It could also be only released to the NHS if you test positive. Lots of stuff could be built in without identifying individuals or locations that is helpful in the fight.
True, good point. There are things that could be included without compromising it - I jumped to a bit of a cynical judgement there.For example, when you set up the app, it could ask you to enter: age, sex, ethnicity, first three letters of home postcode, first there letters of work postcode. Those could all be optional but would help track demographics and hotspots in the virus. It could also be only released to the NHS if you test positive. Lots of stuff could be built in without identifying individuals or locations that is helpful in the fight.
And apologies, I may have been partly responsible for diverting your simple poll thread into more of a discussion!
No worried. I'm as interested in people's reasons why they wouldn't. (I would, as I said in the first post).
Some looks to me like tin-foil hattery. Some sensible concerns. Some luddites. Some practical concerns. Some troglodytes.
It also amuses me that most smart phone users happily give up far more data to the software houses and shopping apps they use than they'd ever give up to the NHS but remain adamant that this is an invasion of privacy.
Some looks to me like tin-foil hattery. Some sensible concerns. Some luddites. Some practical concerns. Some troglodytes.
It also amuses me that most smart phone users happily give up far more data to the software houses and shopping apps they use than they'd ever give up to the NHS but remain adamant that this is an invasion of privacy.
Gassing Station | News, Politics & Economics | Top of Page | What's New | My Stuff