NHS opts for centralised contact-tracing app

Contact-tracing apps are considered an essential step towards safely relaxing lockdown measures, along with widespread testing, social distancing and enhanced hygiene. A contact-tracing app keeps a record of other app users encountered nearby (detected via Bluetooth), such that when an app user tests positive for Covid-19, they can register their infection, triggering an alert to everyone they have recently been close to, advising them to immediately self-isolate and/or get tested.

NHS experts have been working on the app with the National Cyber Security Centre, which has been involved in an advisory role. The app will reportedly be ready for launch “in the coming weeks”.

Matthew Gould, CEO of NHSX, told MPs today that the app will be ready in two to three weeks. It will initially be trialled on a small scale to gauge its success before a national rollout.

The UK's contact-tracing app will send alerts anonymously, so users will not know who has tested positive. NHSX has opted for a centralised model, declining to use the decentralised model being developed by Google and Apple for Android and iOS.

Google and Apple are working together to launch an API in May which enables interoperability between Android and iOS for contact-tracing apps developed by public health bodies. The Google-Apple model involves the exchange of a key code between nearby phones; when a person becomes infected, they consent to sharing their key with the database. The database is regularly downloaded to users’ phones to check for matching codes. Last week, the companies announced that – following consultations – they plan to preserve privacy by having key codes change every 10-20 minutes, making it more difficult for attackers to link a code to a specific user.

With the NHS model, the matching process will happen on a central server rather than on users’ phones. This may spark concerns from privacy activists, who argue that this approach could allow for individuals’ social interactions to be inappropriately tracked. However, a blog post from NHSX CEO Matthew Gould and app manager Dr Geraint Lewis emphasises that the app is being developed with privacy and security as a priority.

They wrote that users could choose to consent to provide more information about themselves in order to provide the NHS with valuable insights into the spread of Covid-19. This could potentially allow for the NHS to send customised messages to users based on their health profiles.

Speaking to the BBC, Professor Christophe Fraser, a University of Oxford academic who is advising NHSX on the app, commented on the centralised system approach: “One of the advantages is that it’s easier to audit the system and adapt it more quickly as scientific evidence accumulates. The principal aim is to give notifications to people who are most at risk of having got infected and not to people who are much lower risk. It’s probably easier to do that with a centralised system.”

The choice of a centralised model sees the UK diverging from the approach favoured by several other European countries, including Switzerland, Estonia, Austria and (following a U-turn) Germany, which are supporting decentralised designs. France remains the only other European country to strongly support a centralised design, although it is facing strong pushback from its own security experts.

The European Commission has indicated that either model is acceptable, although it has warned that privacy must be respected. There are two main European research groups working on contact-tracing protocols: the PEPP-PT consortium (to which Professor Fraser belongs and which supports centralised solutions) and DP3T, which is working on a decentralised protocol with a focus on privacy.

UCL academic and DP3T member Dr Michael Veale commented: “All countries deploying an app must put adoption at the front of their mind and if it doesn’t work well or significantly depletes battery life, then that may act as a deterrent, particularly for those with older phones.”

There are concerns that some contact-tracing apps which do not make use of the Google-Apple toolkit may end up running down users’ batteries too quickly. The proposed NHS app would briefly 'wake up' the app in the background every time the phone detects another device running the same app, adding to the drain of the device's battery. Apple’s solution, on the other hand, is designed to minimise power consumption by allowing matching to be done without the app running, as well as keeping the Bluetooth identifying 'handshakes' between two devices as short as possible.

The NHS has said that it has found a way to make the app work “sufficiently well” on iOS without users having to keep it active.

“Engineers have met several core challenges for the app to meet public health needs and support detection of contact events sufficiently well - including when the app is in the background - without excessively affecting battery life,” an NHSX spokesperson said.

This issue has reportedly posed problems for contact-tracing apps in other countries. Australia – which recently released a contact-tracing app which does not use Apple’s contact-tracing toolkit – has acknowledged power consumption issues, among other problems.

Research indicates that a minimum of approximately 60 per cent of a population must use a contact-tracing app in order for it to be properly effective. As a result, public perception of how well any app works – from privacy protection to power efficiency – are likely to be critical to the success or failure of these apps.

This article was edited on 29 April 2020. It previously stated that Faculty and Palantir were involved in developing the NHS contact tracing app; these are among the tech companies working with the NHS on the Covid-19 Datastore but not known to be working on the contact tracing app.