Epic has made changes to its sepsis prediction model in a bid to improve its accuracy and make its alerts more meaningful to clinicians.
An Epic spokesperson told Becker's in an emailed statement that it began the development of its new sepsis predictive model in February 2021 and released it to customers in August.
The upgrade, according to Epic, was made to improve the software.
"As we develop new tools, we identify opportunities to use them to better serve our customers," the Epic spokesperson told Becker's.
Epic has also changed its definition of sepsis to match the international consensus definition for sepsis.
"One of the most challenging aspects of sepsis is that it doesn't have a single, universally accepted definition," the Epic spokesperson wrote. "Sepsis-3 (the definition that we now use) didn't exist when we developed our first sepsis model, and other definitions continue to be evaluated by industry experts. That said, Sepsis-3 is a current international consensus definition for sepsis. Doctors from leading healthcare organizations across the country helped us determine that it's also the best definition to use for our new predictive model."
The upgrade to the software comes after a study published in JAMA Internal Medicine in June 2021 criticized the sepsis model.
Researchers used data from nearly 30,000 patients in University of Michigan hospitals and found that the sepsis model performed poorly.
According to the study, Epic's algorithm missed two-thirds of sepsis cases, rarely found cases medical staff did not notice and frequently issued false alarms.
When asked about the false alarms, the Epic spokesperson said their customers have also had the ability to tailor the sepsis model to their specific practice.
"Sepsis is complicated, but this model and others like it can help clinicians make timely decisions that save lives," the Epic spokesperson wrote.