A new smartphone app designed to slash the risk of people misreading rapid Covid test results has been released.
The AI-powered tech comes as 57million Covid test packs have been sent to schools ahead of the reopening in England on Monday.
French researchers said up to one in five rapid Covid tests produced difficult to read pregnancy test style bands.
They hoped their new xRcovid app can help boost the accuracy of the “highly subjective” readings set to take place in schools.
It could also help with rapid diagnostic tests (RDTs) used in Covid drive-thru centres, which give results between five and 20 minutes.
And help with testing other conditions where the tests are used including pregnancy, malaria, antibiotic resistant “superbugs” and even Disease X which experts fear could spark a pandemic worse than Covid.
The researchers’ tests using iPhone cameras found the machine-learning app had a 99.3% precision versus reading by eye.
Their paper published in PNAS journal said: “Using the app replaces the uncertainty from visual RDT interpretation with a smaller uncertainty of the image classifier, thereby increasing confidence of clinicians and laboratory staff when using RDTs, and creating opportunities for patient self-testing.
“Since the beginning of the COVID-19 pandemic, new RDTs for identifying SARS-CoV-2 have rapidly proliferated.
“However, these seemingly easy-to-read tests can be highly subjective, and interpretations of the visible ‘bands’ of colour that appear (or not) in a test window may vary between users, test models, and brands.
“Across 11 COVID-19 RDT models, the app yielded 99.3% precision compared to reading by eye.”
“Out of 3,344 tested samples, only 18 False Negatives were observed and were due to low band contrast.”
The researchers said their app used the “high-resolution imaging capabilities of a smartphone camera with an Artificial Neural Network” (ANN) to interpret test results.
They added: “The xRCovid app’s simple interface displays a clear positive/negative outcome and provides information depending on the result (i.e., care-seeking guidance for positive results, precautionary measures for negative).
“Future iterations may eventually provide results directly to a managing physician. The whole analysis is performed on the device, and full privacy for the diagnostic result is guaranteed.”
The advantages of the tech included: “The app’s timer function facilitates timely reading (usually 7 to 15 min) decreasing the number of False Positives.
“Results are independent of human error and subjectivity (up to 20% of one RDT presented faint, difficult-to-interpret bands; these represent the entirety of the FN samples in our experiments, where the ANN, despite having been trained on images with enhanced contrast, is unable to distinguish the bands from the background).
“The app displays results unambiguously (positive, negative, or invalid) without interpretation, translation errors, or jargon.
“App location data can direct users to local health services for medical advice.
“Finally, use of the app by health authorities using (fully anonymized) location data could produce live disease maps.”