Border officials in New Zealand have begun to trial an app designed to detect Covid-19 before the first symptoms of the disease arise, known as ëlarm. The platform has been developed by artificial intelligence (AI) company Datamine, and links with smart watches and other wearable devices to measure metrics like heart rate, temperature or oxygen saturation.
Datamine claims the app can spot the warning signs of Covid-19 with up to 90% accuracy, up to three days before symptoms appear. The app creates a custom baseline for each user from their wearable data history, then uses artificial intelligence to detect physiological changes that might indicate the user is getting sick before they feel unwell. The platform is device-agnostic, and can run on multiple different pieces of hardware.
Up to 500 border workers can volunteer to participate in the trial, which will run until early May to see how the platform works in real-life settings. As new Covid-19 cases in New Zealand are virtually only seen in arriving international travellers, border staff arguably face the most risk of exposure to the virus, and therefore could stand to benefit the most from ëlarm.
New Zealand Ministry of Health Deputy Director Shayne Hunter said: “If the ëlarm app lives up to its potential, it might provide early notification to our critical border workforce if they’re becoming unwell. That means they can take appropriate action such as self-isolating and being tested for Covid-19.”
Can a platform like ëlarm really work?
Data from several studies suggest that wearable devices really could help to predict the onset of illness before it happens. Researchers at Rockerfeller Neuroscience Institute have reported that data from the Oura ring, a wearable sleep and activity tracker, can be combined with an app that measures vital signs to predict the onset of Covid-19 symptoms in advance. They found that the device successfully predicted symptoms like cough, fever and shortness of breath up to three days before their onset.
Early results from the Scripps Research Translations Institute’s DETECT study have also found that wearable fitness devices could improve public health efforts to control Covid-19. DETECT researchers reported that evaluating changes in metrics like heart rate, sleep and activity levels, along with self-reported symptom data, can help identify cases with greater success than just looking at symptoms alone.
Scripps Research Translational Institute director of artificial intelligence Giorgio Quer said: “Early identification of those who are pre-symptomatic or even asymptomatic would be especially valuable, as people may potentially be even more infectious during this period. That’s the ultimate goal.”
The impact of wearable device data could go far beyond individual cases
A study published in January 2020 in The Lancet Digital Health evaluated the use of resting heart rate and sleep data harvested from wearable devices to improve state-level surveillance of influenza-like illnesses (ILI) in the US. Using de-identified sensor data from 200,000 Fitbit users, the researchers found a strong correlation between abnormal data metrics and week-to-week ILI case rates.
They argued that this information could be vital to enact timely outbreak response measures to prevent further transmission of influenza cases during seasonal outbreaks. If state bodies are given access to these metrics, they might be able to spot outbreaks of infectious respiratory diseases – including Covid-19 – before they can spread too far.
Of course, implementing something like this as a public health policy carries with it the usual questions around patient data privacy. It would be vital, if such a scheme were to be implemented, that the data would be completely anonymous and used only with the wearable user’s consent. But if appropriate conditions can be met here, and trials into technologies like ëlarm continue to deliver positive results, then wearables may have a significant role to play in the future of infection control.