Fig. 1
From: Projecting COVID-19 disease severity in cancer patients using purposefully-designed machine learning
Overview of the study. a Data for 348 inpatients at Memorial Sloan Kettering Cancer Center were analyzed. For each patient, up to 267 clinical variables were collected, including basic patient information, cancer history, ICD medical history, laboratory work, and radiology work. Variables were only collected up to the patient’s COVID-19+ date (time zero). b Variables are inputted into a machine learning algorithm (a random forest classifier), which learns to predict patient outcomes based on interactions between multiple variables. c Three possible patient outcomes. Of the 348 patients, 206 did not require high levels of oxygen support, 71 required oxygen support within 3 days of being tested positive for COVID-19, and 71 patients required oxygen support after 3 days