Researchers at Evanston, Ill.-based Northwestern University used machine learning tools to analyze EHR data and find that secondary bacterial pneumonia was the major cause of death in seriously ill COVID-19 patients.
The study analyzed data from 585 patients in the intensive care unit at Northwestern Memorial Hospital in Chicago who had severe pneumonia and respiratory failure. Of those patients, 190 had COVID-19. Researchers used a new machine learning approach called CarpeDiem, which groups similar ICU patient days into clinical states, according to a May 5 Northwestern news release.
The research casts doubt upon the cytokine storm, an overwhelming inflammatory response that leads to organ failure, as the cause of death in COVID-19 patients.
"The application of machine learning and artificial intelligence to clinical data can be used to develop better ways to treat diseases like COVID-19 and to assist ICU physicians managing these patients," Catherine Gao, MD, Northwestern physician and study co-author, said in the news release.