World first study used AI tool to predict COVID patients’ oxygen needs

In a groundbreaking study, researchers employed an AI algorithm to analyze chest X-rays and anonymized electronic health records from COVID-19 patients across 20 hospitals worldwide, including Addenbrooke’s Hospital in Cambridge. The AI tool, named EXAM, accurately predicted patients’ oxygen requirements within 24 hours of emergency department admission, achieving a sensitivity of 95% and specificity over 88%. This advancement enables healthcare providers to anticipate and address severe respiratory complications in COVID-19 patients more effectively, potentially improving patient outcomes and resource allocation during the pandemic.