A tale of two smartphone apps

This post was written on May 10th 2020.

This month's T-SQL Tuesday is hosted by Glenn Berry (b|t). Glenn's invitation was “to write about what you have been doing as a response to COVID-19”, and it feels like my answer should be this: not much. I've stayed home; I've reduced our grocery shopping to once every 7-10 days; I've kept 2m (6ft) away from people outside the house, while maintaining closer virtual contact with family and friends.

As the situation developed in the UK we put cards with our contact details in our neighbours' letterboxes, in case they needed anything (they haven't). I signed up as an NHS Volunteer Responder, which has required me to do… absolutely nothing. In this post I look at the volunteer scheme's GoodSAM smartphone app to see if that could explain why, then talk about an app intended to support COVID-19 contact tracing.

The NHS Volunteer Responder programme was launched on March 24th 2020 by UK Health Secretary Matt Hancock. The intention of recruiting volunteers was to reduce pressure on the National Health Service (NHS) by delegating “simple but vital tasks”. These could be delivering medicines to patients unable to leave their homes, doing their shopping, or just giving them a phone call to reduce their social isolation.

I was among 750 000 people to sign up and download the GoodSAM smartphone app in the five days following the launch. In the app (screenshot on the right), users declare themselves “on duty” in order to receive notifications of allocated tasks. At the time of writing I've received precisely zero notifications, despite (by the app's calculation) 127 hours on duty. So what's going wrong?

At this year's virtual GitHub Satellite conference I watched a panel session on the subject of Open source for good – COVID-19 pandemic responses (3h55m in if you want to check it out). One of the panellists, Linda Raftree (b|t) described three questions we should be asking ourselves when bringing technology to bear in a crisis:

1. Does it solve a real problem? Have we understood the problem? Is it a problem that technology can solve? Is the technology appropriate?
2. Does it link to the wider ecosystem? Not just the question of technical interoperability, but how does the technology sit within the health, social, economic and legal contexts? Does it build on what already exists?
3. Is there potential for harm? Other than us, who's bearing the risk? What are the longer-term implications of a short-term fix?

This got me reflecting on the use of GoodSAM to manage NHS Volunteer Responders.

The GoodSAM app predates COVID by five years. Its original purpose was to notify off-duty healthcare professionals of nearby cardiac arrests, to improve survival by getting clinicians to patients even before emergency services arrive. The idea of “crowdsourced resuscitation” was rolled out to crowdsource NHS Responder Volunteers. So…

• Does it solve a real problem? Volunteer engagement hasn't required strenuous effort (the programme paused accepting new volunteers after just five days and at the time of writing hasn't started again). Distribution of tasks doesn't seem to be overwhelming – I'm not alone in reporting having nothing to do.
• Does it link to the wider ecosystem? The fact that GoodSAM is an existing app means that it's already well-supported and well-understood. Tick!
• Is there potential for harm? It's clear that COVID will be here for a while, so even if volunteers aren't needed now, demand may change – but with wannabe responders expressing frustration I wonder about disengagement creeping in before the real work arrives. The Association of Directors of Adult Social Services criticised the national recruitment drive for having “diverted 750,000 volunteers away from supporting local communities.”

While the choice of app makes good use of an existing resource, my answers suggest it may not be the best fit for this scenario – and perhaps that contributes to my experience and that of other volunteers.

Another piece of COVID-related smartphone software making regular appearances in UK news is the proposed NHS COVID-19 app, currently in development by NHSX (the technology arm of the NHS). Contact tracing is expected to become an important part of slowing the spread of COVID as movement restrictions are eased, and the new app is intended to automate much of the process. Those questions again:

• Does it solve a real problem? The problem is real enough, but the app can't solve it on its own. Bluetooth proximity sensing can be error-prone, and smartphones are an imperfect proxy for human contact detection – my phone might be “near” my neighbour's, even though there's a wall between us. On April 23rd the UK government announced its intention to recruit 18 000 “contact tracing agents” to work alongside the app.
• Does it link to the wider ecosystem? The app has been criticised for a centralised model that is less protective of users' privacy than an alternative developed by Apple and Google, and which may not be legal. Latest reports at the time of writing suggest that the UK government could be considering a change of approach.
• Is there potential for harm? The question of legality alone suggests potential for harm, indeed numerous UK academics have expressed concern that contact tracing apps create disturbing precedents for state surveillance and risk mission creep.

In this case too it seems that these questions have yet to be answered fully.

As technologists we're fascinated by solving engineering problems, but sometimes we're less good at working out how – or if – a human problem can be mitigated with a technological solution. When you have a hammer (or in this case a smartphone app), suddenly everything looks like a nail.

The thing I love about these simple questions is that they seem to have no easy answers – so they're probably great questions! I'll be posing them to myself a lot more in the future.