On Tuesday 28 April 2020 in a session with MPs on the Commons science and technology committee, Matthew Gould, CEO, NHSX said the first version of the contact tracing app will be ready in two to three weeks.
Gould said the app will be trialled in a small area, and if it performed well, it will be ready when the country is looking to ease lockdown measures.
Work started on the app two weeks before the start of the lockdown, on 7th March, with Gould commenting yesterday “I hope it will be ready when it is most needed when the country is looking at tools to ease lockdown safely.”
The app uses bluetooth technology, so if you become unwell with the symptoms of coronavirus, you can securely tell the NHS app and the app will then send an alert anonymously to other app users that you have been in significant contact with over the past few days, even before you had symptoms, so that they know and can act accordingly.
When questioned extensively on how the data might be used during and after the pandemic and on privacy, Gould re-iterated that all data will only be used for NHS care and research. For research users would have to opt-in.
Professor Christopher Fraser, Senior Group Leader in Pathogen Dynamics at University of Oxford Big Data Institute commented in the session on manual contact tracing and the effectiveness of using technology. He said “For manual contact tracing the rule is you ask someone if they have been in contact with someone for 15 minutes and within two metres, and of course that can be evaluated. The Bluetooth model is a continuous model. The challenge is turning that into a binary decision. Was that a contact that was meaningful that transmission could have happened.”
Gould said “On one level there will always be a level of binary quality to it…are you being advised to isolate or are you not being advised to isolate. Behind that there can be a more subtle algorithm.”
“With a powerful computer in people’s pockets it can work on a more subtle contact risk model, for example you may have had one significant contact over a prolonged period with someone who subsequently became symptomatic or you may have had a series of shorter contacts, both of which contribute to the level of risk.”
Gould confirmed the commitment to publish the risk model so its transparent. He said “We will open source the software and publish the privacy model. The whole model exists on people having randomised/anonymised IDs. I do believe what we have done is respectful of privacy.”
The unit has also been cooperating closely with a range of other countries, sharing code and technical solutions.
