Dr. Brian Eastridge has spent much of his career as an Army trauma surgeon in war zones, and that experience shapes how he sees his current job as chief of the Division of Trauma and Emergency Surgery at UT Health San Antonio.
"We tend to meet people on the worst days of their lives," Eastridge said. "It's very, very high stakes."
Now Eastridge is applying that same urgency to a different problem: patients in rural Texas who need advanced trauma care aren't getting it fast enough.
Trauma is the leading cause of death for people up to age 45. In Texas, state EMS and trauma registry data shows up to 80% of the most severely injured patients wait more than two hours to be transferred to a trauma center.
Eastridge co-leads iRemedyACT, a statewide research consortium using artificial intelligence to close that gap. He walks through the chain of events that creates the delay: An EMS crew picks up a patient and takes them to the nearest facility, often a Level Four trauma center, essentially a well-equipped rural emergency room without dedicated trauma surgeons on call.
"That doctor has to evaluate the patient and make a decision," Eastridge said. Then, if a transfer is needed, "they have to hospital shop to find a trauma center that will take the patient" and call an ambulance to carry it out. Rural ER doctors treat everything from strokes to pediatric flu, he noted, and trauma isn't usually their specialty. "Trauma is generally not one of their specific strong sets, because generally these rural communities don't see a lot of injury."
That's where AI comes in, Eastridge said. It wouldn’t replace the doctor's judgment, he explained, but sharpen it. The tool his team is developing flags severe injury patterns and prompts an earlier transfer call. "It pushes the doctor to make that decision and make that transfer call earlier," he said, "rather than trying to specifically identify injury patterns or specific injuries prior to the patient's transfer."
This could cut delays by a significant amount, Eastridge said. Transfer decision could be made in five to 10 minutes, though the current standard is 15 to 30. Closing that gap, he believes, could save lives. “Particularly for the severely injured patients,” he said. “We suspect there would be a significant survival benefit.”
Eastridge acknowledges the unease many feel about AI making life-or-death recommendations. "It is just a tool," he said. "You don't use your ultrasound and exclusively make your decisions with that ultrasound. I think we should consider this today as just another tool in our clinical toolbox to actually get at making the right decision for the patient."
With supplemental funding from the UT System's AI Beacon Moonshot program, Eastridge said the team plans to launch a statewide pilot of the decision-support tool this fall.
Science & Medicine is a collaboration between TPR and UT Health San Antonio about how scientific discovery in San Antonio advances the way medicine is practiced everywhere.
