The New York University College of Dentistry created a mobile app that uses artificial intelligence to analyze risk factors and biomarkers from blood tests to produce a COVID-19 "severity score" for patients.
The app, which predicts how severe a COVID-19 patient's case will be, has been evaluated within the Family Health Centers at NYU Langone Health in New York City. The researchers are looking to launch it nationwide in the next few weeks, and COVID-19 severity scores derived from the app could be integrated with hospital EHRs, allowing clinicians to use the assessment tool to support care decisions.
To build the app, the NYU researchers used data from 160 hospitalized patients in Wuhan, China, to pinpoint four biomarkers measured in blood tests that were significantly higher in patients who died than those who recovered. The biomarkers are C-reactive protein, myoglobin, procalcitonin and cardiac troponin I.
Researchers then used the biomarkers and two established risk factors of age and sex to build a machine learning model. When a patient's biomarkers and risk factors are entered into the AI-powered model, it generates a numerical COVID-19 severity score ranging from zero, mild or moderate to 100, which is critical.
The team used data from 12 hospitalized COVID-19 patients in Shenzhen, China, to validate the model and then further validated it by testing it with data from more than 1,000 New York City COVID-19 patients.