Just a few artificial intelligence models have received FDA emergency use authorization to help clinicians diagnose and treat COVID-19, the latest being an algorithm designed by Oakland, Calif.-based digital health company Dascena.
Four details about the algorithm:
- The algorithm, named the COViage Hemodynamic Instability and Respiratory Decompensation Prediction System, predicts the likelihood hospitalized COVID-19 patients will need intubation.
- During spring, Dascena tested the AI tool in a clinical trial involving 197 hospitalized COVID-19 patients across five U.S. hospitals. The study found that the algorithm performed more effectively than the Modified Early Warning Score, a rules-based system that is not AI-reliant but is commonly used by clinicians to help assess COVID-19 patients' needs, according to STAT.
- The FDA granted emergency use authorization to the algorithm Sept. 24.
- In the letter the FDA wrote to establish the algorithm's emergency approval, it warns that the tool is not "intended to be used on a standalone basis for clinical decision-making."
Pittsburgh Health Data Alliance, Amazon Web Services creating AI models to detect breast cancer, depression more quickly