Smartphone-based COVID-19 detection take a look at from Australia reveals excessive accuracy


Share post:

ASX-listed digital well being agency ResApp Well being has introduced that its pilot scientific trial for a smartphone-based COVID-19 screening take a look at has achieved constructive outcomes. 

The brand new diagnostic app, which makes use of machine studying to analyse the sound of a affected person’s cough, was reported to appropriately detect COVID-19 in 92% of contributors with the an infection through the trial. With a 92% sensitivity, it exceeds real-world measured sensitivity of speedy antigen assessments, ResApp claims. 


The scientific trial in the US and India concerned 741 sufferers with round 60% of them COVID-19 constructive. 

Primarily based on preliminary outcomes, the app was additionally discovered to be “excellent,” having attained an space beneath the curve of 0.9, when measured for the way effectively it distinguishes between two diagnostic teams. It additionally obtained 80% specificity in figuring out sufferers who examined destructive for COVID-19. 

To make sure that its algorithms are detecting COVID-19, it was checked towards the Breathe Straightforward dataset, which incorporates information from over 1,000 sufferers with varied non-COVID-19 respiratory situations. 

ResApp additionally shared that the app confirmed “constant efficiency” over two durations, one when Delta was the key variant and the opposite when Omicron was the dominant variant.

The corporate additional famous that the app demonstrated “decrease” efficiency in asymptomatic sufferers, though it solely recruited a small variety of them.

ResApp stated it should publish the outcomes of its scientific trial in a peer-reviewed journal “within the coming weeks”.


The World Well being Group has warned that the worldwide pandemic is way from over and that well being techniques have to be ready to face extra harmful variants.

ResApp goals to initially convey its newest screening resolution in locations the place frequent COVID-19 testing is required, reminiscent of workplaces, hospitals, faculties, journey areas, sports activities venues, leisure hubs, and aged care centres. With its smartphone-based take a look at, speedy antigen or PCR assessments can be required much less to substantiate COVID-19 in these settings given its excessive sensitivity and specificity.

ResApp CEO Tony Keating stated they’re now reaching out to regulators throughout markets and have already engaged with international well being tech corporations to roll out its app-based screening take a look at. 

Furthermore, the COVID-19 screening app additionally has the power to “enhance safety and reporting of outcomes” utilizing biometric identification like facial recognition.


Different tech builders have constructed algorithms to detect COVID-19 by analysing sufferers’ cough sounds. In 2020, researchers from the Massachusetts Institute of Expertise within the US developed an AI software to diagnose COVID-19 that was skilled utilizing over 70,000 audio recordings, together with these from contaminated sufferers. In October final 12 months, Andhra Medical Faculty in India tried out Salcit Applied sciences’ Swaasa AI platform, which conducts an audiometric evaluation of cough sounds, together with temperature, oxygen saturation, signs, of suspected COVID-19 sufferers.

In the meantime, Breathonix, a spin-off firm from the Nationwide College of Singapore, has developed a breath take a look at that may quickly detect COVID-19 inside a minute. 


“By quickly ruling out COVID-19, ResApp’s COVID-19 take a look at would considerably cut back the variety of speedy antigen and PCR assessments required, whereas nonetheless sustaining the illness surveillance wanted to handle the continued affect of COVID-19. The simplicity, ease of use and limitless scalability of ResApp’s take a look at shall be welcomed by public well being officers world wide,” Catherine Bennett, a professor and the chair of Epidemiology at Deakin College, commented.

Supply hyperlink


Please enter your comment!
Please enter your name here


Related articles