COVID-19: WHICH PARTS OF THE NHS COVID-19 APP ARE WORTH USING?

The NHS COVID-19 app was launched with great fanfare for England and Wales just over a week ago. At least 14.5 million downloads from a population of around 60 million.  Some 24% of the population, enough to potentially be really valuable in the fight against COVID-19.  But only if people actually use it, and only if the app works 'sensibly'.

This is an update to the article here.  I said I had hoped to use the app, but would not download it until a number of concerns had been overcome. This was because of initial concerns it had been poorly designed and tested, and appeared to not be fit for purpose.

A few days later and some new issues have arisen, and also that some aspects are not as serious as first appeared:

• On the negative side, for example, 'phantom alerts' outside of the actual app are confusing people, suggesting little if any 'beta testing' has taken place with end-users.  Some beta-testing has taken place, but clearly not enough. What else is going wrong in the app?

• On the positive side, the authorities can't know (at least as yet) if you have been sent a self-isolate alert.  So there is no risk of a £1000 fine, as is the case with an instruction from a human contact tracer

• Also alerts to self-isolate do not occur as a result of scanning into a venue where someone else has been who has tested positive, neither automatically nor manually.  I had initially assumed self-isolation alerts would happen, because without them there is little value of venue scanning in the fight against COVID-19.  All that happens is you could get an alert to watch out for symptoms.  But people will be doing that anyway.  Presumably it would be useful for human contact tracers to follow up on venues where infected people have visited.
  

So my fundamental conclusion remains unchanged.  Having worked extensively in the development and testing of much more sophisticated systems, the app comes across as having been poorly designed and inadequately tested.  Not fit for purpose.  In particular over-zealous in alerting people to the risk of having been infected.  

Though the app could offer some advantages from using it.  Confused?  Let me explain.

Interesting that the Police force across the country have been instructed not to load the app on to their work phones, and not to use certain aspects of the app on their personal phones.  That could be because:

• The Police authorities share my concerns
• And/or have found more concerns.  
• Or they are erring on the side of caution as they simply don't know, because the app was launched before the Police had been involved in the testing.  In which case that would underline my concerns about inadequate testing, and the resulting lack of trust in the app.
  

THE BASIC TECHNOLOGY

This is the second iteration of the app, with a totally different de-centralised data architecture and proximity engine than the first version.  There have been millions of downloads despite:

• Only certain phones are suitable for download.  A smaller number than the first app
• Not being possible to download to tablets

The tablet restriction is odd.  I often take my iPad when I go to the pub, typically to look things up and show it to my pals on a larger screen than my phone.  It would be more useful to scan the QR code and see the pub or restaurant's menu on the larger screen.  My iPad has a valid iOS release, yet the app won't download.  I've tried it.

But my mobile phone is Android (I like to keep a foot in both camps).  Will I download the app onto there?

Possibly now.  Not because the app is much good, but the risks of not using the app could be greater, given the potential £1000 fine  Certainly now I am getting emails that various venues that I visit are displaying the QR code, but where there is little or no chance of being close enough to someone to be infected.

SO WHY MIGHT I USE THE APP?

In principle I was hoping that the app would be good enough to download and use.  But it looks as if it was rushed out and released before it had been adequately "beta tested" with members of the general public, the police and others.  Anybody downloading it is effectively taking part in a giant beta test.  The results so far are not satisfactory.  

The main function of the app is to help to reduce infections.  That is by advising users when they have been potentially exposed to infection, in the hope those people will self-isolate before they show symptoms, thereby reducing transmission of the virus:

• An alert advising self-isolation according to the proximity algorithm
• An alert to watch out for symptoms if used a QR code to check into a venue.  Frankly that adds nothing but worry, as anyone sensible would be looking out for symptoms anyway
  

In each case, the central system, and the authorities using it, cannot see who has been alerted (although this might change), and so it has been confirmed that no fines can be levied.

The BBC have published an article confirming some problems with the app, notably relating to not being able to check out of a venue.  What is said is "you cannot be alerted that somebody arrived in the pub hours after you left, and they later tested positive, because your phones could not have been physically close to each other at the same time".  That is true, in terms of a self-isolation alert.  Whether you have the proximity function in the app switched on or not.  But you can bee still notified by the QR code alert type, because your phone and the central system don't know when you left that venue that day, only if or when you check into another.

£1000 fines can however be levied if they know who you are.  Such as if a venue is asked for a list of visitors who have provided their details to the venue because you haven't used the app to check in.

Another issue is that the app is known to consume significant battery power,:

• It needs to have Bluetooth switched on for the proximity sensor to work
• The app itself uses power, with or without Bluetooth 

Just to confirm there is no compulsion to use the app if you have downloaded it.  

 

SO HOW MIGHT I USE THE APP?

 

I only use Bluetooth for my phone in the car, where I can keep it powered from the cigar lighter.  Otherwise experience tells me the phone battery can die if too much is running.

So I won't be switching Bluetooth on in a pub garden, for example. Nor indoors, in a pub, at home, nor anywhere else.  That also means I won't be plagued by false self-isolation requests, which is my principle concern. 

But if I arrive at a venue which is legally compelled to display a QR code for me to scan, scanning it would avoid the alternatives if I didn't scan it.  I am supposed to:

• Register with the venue through the internet onto an electronic list, or
  
• Register on a manually prepared list

In both cases those lists could be made available to public health personnel who could then compel me to self-isolate.  I would risk a £1000 fine if I didn't.  Even if I know I was nowhere near anybody to possibly catch the disease.

 

So I could switch on the app to check in to a venue, and switch it straight back off again, without at any stage switching on Bluetooth.

 

Just to reinforce a point.  I am perfectly happy to do my duty and self-isolate, but only if I can be confident that there is a significant risk of me having caught the disease.  The app's design and limited testing gives me no confidence that is the case, and the inherent risk of false alerts is too high.  So I am not prepared to use it, other than reducing the risk of my bearing a fine or unnecessary loss of earnings, especially as I am self-employed.

 

Having downloaded the app, there are various other useful functions, but all can be done without the app.  Though perhaps more conveniently with it.

 

Update 12/10/20:  The App will have details of the local restrictions.  The information will be available on the main web sites, but will be especially useful on the app.

 

IN CONCLUSION

 

I might (but only might) now download the app on my phone for two reasons:

• Check into venues, to avoid the alternatives
  
• Convenience of the functions not involved with alerts
  

I will not be using the app to use the proximity sensor.  That is to avoid false self-isolation alerts.