Researchers from an expert advisory group formed by Science Foundation Ireland (SFI) have been drafted in to conduct an independent assessment of the effectiveness of aspects of the State's Covid-19 contact tracing app.
Dr Brendan Jennings, director of CONNECT, the SFI research centre for future networks and communications, said the group will particularly focus on the accuracy of how the app uses Bluetooth technology to estimate when someone has been in close contact with a person infected with coronavirus.
The accuracy of the Bluetooth technology has been called into question, giving rise to questions about how it would function in crowded areas.
“We’re going to do some testing around the app and do an independent assessment from that particular part of the functionality, the accuracy of estimation of distance,” Dr Jennings said.
It follows on from research undertaken by scientists in Trinity College Dublin, who are also involved in the group, which called into question the accuracy of the app. They found that it was likely to be challenging to use Bluetooth technology to reliably detect when people using the app were within two metres of each other.
“[The research will] basically try to replicate real-use cases – where people over the next couple of months will be interacting, in supermarket queues or in the supermarket itself, on public transport when you might end up close to people you don’t really know.”
The group has been working with the HSE already, but Dr Jennings said the intention was to help “fine-tune” the app, which has now been developed and is being put out for testing this week.
Garda
Initially, the testing will be done by asking members of the Garda to install the app on their phones and carry them while circulating on duty. This is designed to test the exposure notification service developed by Apple and Google, which is used in the Irish app.
“That’s where the app is going to be most useful; we want to replicate those scenarios to the degree we can, and estimate the accuracy of the distance between two phones,” Dr Jennings said.
The HSE said its team is “working closely” with SFI “to access expertise across a number of fields including the use of Bluetooth technology. This has helped to inform the development of the app to date and will assist us to address any issues highlighted during the final test.”
The app was originally scheduled for launch in early April, but that date has been continuously pushed back as technical issues and privacy concerns emerged. Some of these issues have been addressed, such as the commitment not to hold data on a centralised Government database, while the HSE will also be using platforms provided by Google and Apple to overcome issues around the app’s functionality on the most popular smartphone operating systems.
However, privacy campaigners say they are withholding their verdict on the app until they see its source code, and the data protection impact assessment (DPIA) which accompanies it. This is a key document that outlines the HSE’s assessment on how data privacy might be impinged on by the project.
Centralised model
Antoin Ó Lachtnain, a director of data protection advocacy group Digital Rights Ireland, praised the HSE for dropping its centralised model, but said "there are still many unanswered questions about the app".
“In particular, the DPIA and the source code for the system have not been published, despite many promises. The system has now gone into a trial phase with human subjects without these steps having been taken. This could undermine the credibility of the whole project.”
The DPIA has been sent to the Data Protection Commissioner as part of a consultative process, but the watchdog has no official role in signing off on the app.
Minister for Health Simon Harris has said at least 25 per cent of the population will need to download the app in order for it to have sufficient coverage to work.
However, there are indications the uptake will have to be higher for the app to be effective.
Oxford University's Big Data Institute has estimated about 80 per cent of smartphone users would need to have the app installed for the system to work properly.
In Singapore, about one million people (about 20 per cent of the population) had downloaded the app by the end of April. At that time, an article in Nature magazine calculated that, in an encounter between two randomly chosen people, there was only a 4 per cent chance that both would have the app.
