Using a pinprick test, the development could predict infection before a standard test can diagnose the presence of the virus, as well as any new version of pathogenic coronaviruses in the future.
Using data which was made public from a hospital in Sao Paulo, Brazil, researchers found they were able to predict who had the virus based on full blood counts. Patients with coronavirus tended to have decreased levels of platelets and types of white blood cells. A simple equation was then devised using four parameters of the blood counts, which could be used in hospitals after validation.
The team behind the project, Modelling and Prediction Pandemics (MaPP), recently won first prize at the week-long international Coronahack, an event which brought together some of the brightest minds in biomedicine and AI to stimulate the prospect of new and novel ideas to combat COVID-19.
Since winning the competition, a paper has been written, submitted and accepted for publication in the International Immunopharmacology Journal in less than 6 weeks, time scale which is far less than similar pieces of research would usually take. The prize money and GPU will be used to continue the research and further develop the AI modelling to help screen during future waves of the pandemic in Gabon.