Every year in the U.S., about 6% of emergency room visits are due to sepsis, an extreme immune system reaction to infection that can lead to organ failure and death in a short period of time. Anywhere from 30% to 50% of hospitalizations that end up in death can be traced to sepsis. But doctors have few tools to predict which patients should be tested for sepsis, since sepsis shares many symptoms with other conditions.
Bobby Reddy Jr.’s dedication to solving this problem extends back 15 years, and became even more urgent when he learned that even if the infection that caused sepsis can be controlled with antibiotics, the resulting damage from an out-of-control immune system couldn’t be reversed. “It became one of our core missions to help these patients never to fall off that cliff,” he says. His company’s new test, Sepsis ImmunoScore, is an AI-based assessment of 22 different factors in a patient’s health records and measurements of vital signs that can accurately predict which patients are most likely to go on to develop severe sepsis, and which patients’ immune systems are less likely to overreact. The resulting score helps doctors decide which people to prioritize for sepsis treatments.