Artificial intelligence has the potential to transform medical care as we know it — but experts have concerns
The year was 2018, and Dr. Gabriel Wardi saw a potential solution to a long-running problem in healthcare: sepsis. Wardi, the medical director for hospital sepsis at UC San Diego Health, says sepsis — an overzealous immune system response to an infection — kills about 10 million people a year worldwide, including 350,000 people in the United States.
Part of the problem with sepsis is that there are a lot of ways it can present, which makes it tricky to diagnose. For years, Wardi had been trying to see if electronic health records could trigger an alert for doctors and nurses when someone becomes at risk.
“Unfortunately, those early alerts were wrong almost all the time, and you can imagine that in a busy hospital, your initial reaction is, ‘Get this thing away from me,’ because it’s wrong all the time, it changes your workflow, and no one likes it,” he says.
But when artificial intelligence entered the scene, Wardi wondered if AI models could more accurately predict who’s going to get sepsis.
