COVID-19 Algorithm Gets FDA Emergency Use Authorization
October 14, 2020
An algorithm developed by Dascena, Inc., a California-based machine learning diagnostic company, has been found to successfully predict which COVID-19 patients are at risk for either unstable blood pressure or respiratory failure. As a result, the Food and Drug Administration has issued the company an Emergency Use Authorization (EUA) for its “COVID-19 Hemodynamic Instability and Respiratory Decompensation Prediction System.”
While the FDA’s EUA for the company’s “COViage” system is not a full approval or clearance, the system, which makes predictions with data taken from patient electronic health records (EHR), can now be used in clinical practice.
“COViage demonstrated the ability to help diagnose respiratory decompensation and hemodynamic instability earlier and more accurately that the standard of care,” said Ritankar Das, president and CEO of Dascena. “We are excited to bring this machine learning algorithm to the bedside, which may enable the preservation of many lives and improve allocation of hospital personnel.”
The algorithm makes use of EHR data (e.g, age, heart rate, temperature, and blood pressure) to calculate patient risk and sends an advance notice to the healthcare provider. The COViage system then integrates with the hospital’s EMR system to display its predictions.
“Based on FDA’s review of retrospective clinical validation data, FDA has concluded that COViage may be effective for use by HCP as a diagnostic aid to assist with the early identification of adult COVID-19 patients who are likely to be diagnosed with hemodynamic instability or respiratory decompensation, which are common complications associated with COVID-19,” the FDA wrote in a letter to Dacsena authorizing the EUA.
The EUA applies to hospitalized adult patients with confirmed COVID-19. Both FDA and Dascena recognized that COViage could miss a case or issue false positives.
The system was evaluated in a clinical trial in which 197 patients were enrolled who visited a hospital emergency department and were subsequently admitted to one of five U.S. hospitals between March 24, 2020 and May 4, 2020. Patients on the trial had their first vital signs taken, and laboratory data available, within two hours of their admission or ED arrival.
“The early identification of patients at risk of respiratory decompensation or hemodynamic instability would enable physicians to more aggressively monitor these patients in a controlled environment and provide earlier treatment,” Das said.
The data from the trial were published in the journal Computers in Biology and Medicine.